diff --git a/Projeto_Final_MoA_no_Paredawn.ipynb b/Projeto_Final_MoA_no_Paredawn.ipynb
new file mode 100644
index 0000000..45c3748
--- /dev/null
+++ b/Projeto_Final_MoA_no_Paredawn.ipynb
@@ -0,0 +1,11475 @@
+{
+ "nbformat": 4,
+ "nbformat_minor": 0,
+ "metadata": {
+ "colab": {
+ "name": "Projeto_Final - MoA no Paredawn.ipynb",
+ "provenance": [],
+ "collapsed_sections": [
+ "US878rR1wrBq",
+ "XBQ40pnrwwYL",
+ "UAK3kawHwyZZ",
+ "fy8XzwXWw0fq",
+ "v2p_CrKixE0k",
+ "rGjBCv6PLsC-"
+ ],
+ "authorship_tag": "ABX9TyOLyjrHzNZ32W4P0HHD/+z9",
+ "include_colab_link": true
+ },
+ "kernelspec": {
+ "name": "python3",
+ "display_name": "Python 3"
+ },
+ "language_info": {
+ "name": "python"
+ }
+ },
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "view-in-github",
+ "colab_type": "text"
+ },
+ "source": [
+ " ***Projeto Final Imersão Alura 3° edição*** \n",
+ "***"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "0LAuJQsEa94g"
+ },
+ "source": [
+ ""
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "5TOR8N9ui7Qk"
+ },
+ "source": [
+ ""
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "5ZL5iRagjBTc"
+ },
+ "source": [
+ "Olá Mergulhadores !!! \n",
+ "\n",
+ "Você deve estar pensando: `O que é esse treco de MoA`, não é mesmo? se não pensou isso, poxa, eu pensei!! \n",
+ "\n",
+ "Mas fique tranquilo, vou tentar descomplicar isso para você! Vamos colocar esse tal de MoA no `paredão` para **analisar e explorar seus dados**, depois vamos tentar criar uma `maquina preditiva` para **prever** com os dados que temos se ele será **atividado** ou **não**. \n",
+ "\n",
+ "Após isso, vamos criar outra maquina preditiva para prever se o experimento feito será com uso de droga ou uso do controle com um melhor metrica de avalição possível. \n",
+ "\n",
+ "Poxa, só de falar estou animado! Espero que você também esteja assim!\n",
+ "\n",
+ "Então vamos lá!"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "E-w_JjsDk64N"
+ },
+ "source": [
+ "#***Objetivos do Projeto*** #"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Rfmk7O_Jk9YU"
+ },
+ "source": [
+ "
O objetivo desse projeto é o desenvolvimento de duas maquinas preditivas, com o intuito de auxilar na descoberta de novos medicamentos através dos MoAs (Mecanismo de Ação).\n",
+ "\n",
+ "\n",
+ "\n",
+ "1. `Primeiro Objetivo:` Criar uma Maquina preditiva que consiga prever se o experimento irá ativar um ou mais MoA, com um F1-score maior que 80%\n",
+ "2. `Segundo Objetivo:` Criar uma Maquina preditiva que consiga prever se o experimento foi tratado com **droga** ou **com_controle**\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Vk_RARVBlJvE"
+ },
+ "source": [
+ "#***Contextualizando*** #"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "I7W8yPGSlMlD"
+ },
+ "source": [
+ "O [Conective Map](https://clue.io/), é um projeto dentro do MIT e Harvord, o Laboratorio de Ciencia Inovadora aprensenta esse desafio com o objetivo no avanço no desenvolvimento de medicamentos por de algoritmos que consigam prever o MoA."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "PZUucP8AlO0j"
+ },
+ "source": [
+ " Mas o que é MoA? \n",
+ "\n",
+ "
Do inglês MoA significa Mechanisms of Action (Mecanimos de Ação).
\n",
+ "\n",
+ "Antigamente, as drogas eram derivadas de produtos naturais, muitos remedios eram colocamos em uso clinico sem ao menos entender os mecanimos biológicos daquele medicamento.
\n",
+ "\n",
+ "Atualmente, com grandes tecnologias, esse processo de descobrimento de novas drogas passou por uma mudança nas abordagens. Hoje, temos um modelo mais voltado para a compreenção do mecanismo biológico de uma doença, com isso, buscamos identificar um alvo proteíco associado a uma doença e desenvolver uma molecula que possa reagir com essa proteína alvo.
\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "hffcf4mMlPjr"
+ },
+ "source": [
+ " Como determinamos os MoAs de um novo medicamento? \n",
+ "\n",
+ "Uma abordagem é tratar uma amostra de célula humanas com a droga e depois analisar as respostas celulares com algoritmos que consigam identificar padrões conhecidos em grandes bancos de dados genômicos
\n",
+ "\n",
+ "Neste projeto tem um conjunto de dados exclusivo que combina a expressão gênica e os dados de viabilidade celular. Os dados são baseados em uma nova tecnologia que mede simultaneamente (nas mesmas amostras) as respostas das células humanas aos medicamentos em um pool de 100 tipos de células diferentes
\n",
+ "\n",
+ "Em outro conjunto de dados, chamado dados_resuldados.csv , tem as anotações do MoA para mais de 5.000 medicamentos
"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "xZVQuT8Rlbjf"
+ },
+ "source": [
+ "Esse projeto foi desenvolvido na [Imersão dados - 3° edição - Alura](https://www.alura.com.br/imersao-dados)***Importação das Bibliotecas*** "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "8U_MU0mKjTl8"
+ },
+ "source": [
+ "import pandas as pd\n",
+ "import numpy as np\n",
+ "\n",
+ "#Bibliotecas de visualização de dados\n",
+ "import seaborn as sns\n",
+ "import matplotlib\n",
+ "import matplotlib.pyplot as plt\n",
+ "\n",
+ "import warnings\n",
+ "warnings.filterwarnings('ignore')\n",
+ "\n",
+ "#Bibliotecas de Pré-processamento\n",
+ "\n",
+ "from sklearn.preprocessing import LabelEncoder\n",
+ "from sklearn.model_selection import train_test_split\n",
+ "from sklearn.ensemble import ExtraTreesClassifier\n",
+ "from sklearn.utils import resample\n",
+ "from sklearn.feature_selection import SelectFromModel\n",
+ "\n",
+ "#Bibliotecas de Modelo de Machine Learning\n",
+ "from sklearn.linear_model import LogisticRegression\n",
+ "from sklearn.tree import DecisionTreeClassifier\n",
+ "from sklearn.ensemble import RandomForestClassifier\n",
+ "from sklearn.dummy import DummyClassifier\n",
+ "from xgboost import XGBClassifier\n",
+ "\n",
+ "\n",
+ "#Bibliotecas de Avaliação do Modelo\n",
+ "from sklearn.metrics import accuracy_score\n",
+ "from sklearn.metrics import classification_report"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "XBQ40pnrwwYL"
+ },
+ "source": [
+ "#***Carregamento das Bases*** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "UAK3kawHwyZZ"
+ },
+ "source": [
+ "## Carregando a Base Experimentos "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "PbTIA2dkw2yS"
+ },
+ "source": [
+ "df_experimentos = pd.read_csv('https://github.com/alura-cursos/imersaodados3/blob/main/dados/dados_experimentos.zip?raw=true', compression = 'zip')"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "fy8XzwXWw0fq"
+ },
+ "source": [
+ "##Carregando a Base Resultados "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "b-5E8BTuc8Su"
+ },
+ "source": [
+ "df_resultados = pd.read_csv('https://github.com/alura-cursos/imersaodados3/blob/main/dados/dados_resultados.csv?raw=true')"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "WVQ17navxCjM"
+ },
+ "source": [
+ "#***Analise Exploratória*** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "o6jqXqkMxpsz"
+ },
+ "source": [
+ "Então vamos começar a nossa análise e colocar os dados das bases `Experimentos` e `Resultados` no `Paredawn`"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "g-AGUpOsN_ae"
+ },
+ "source": [
+ "Base Experimentos "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "2a2D7ZoOxl2T"
+ },
+ "source": [
+ "Para iniciar a nossa análise exploratória vamos começar entendendo a dimensão na nossa base utilizando o `.shape`:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "CAQj2A4exIJ9",
+ "outputId": "facbb5ae-68ec-42b3-e92d-903a0c983929"
+ },
+ "source": [
+ "df_experimentos.shape"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "(23814, 877)"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 144
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "fPh-ybHVx5u2"
+ },
+ "source": [
+ "Conforme observado acima, conseguimos perceber que temos `23.814 linhas` e `877 colunas` "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "8MtOef0Ux9Lq"
+ },
+ "source": [
+ "Para verificar as primeiras linhas do nosso DataFrame utilizaremos `.head()`"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 244
+ },
+ "id": "LW6K5VPIyJmL",
+ "outputId": "75447055-6851-42d9-ecd3-b5a7c17c70a9"
+ },
+ "source": [
+ "df_experimentos.head()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " id \n",
+ " tratamento \n",
+ " tempo \n",
+ " dose \n",
+ " droga \n",
+ " g-0 \n",
+ " g-1 \n",
+ " g-2 \n",
+ " g-3 \n",
+ " g-4 \n",
+ " g-5 \n",
+ " g-6 \n",
+ " g-7 \n",
+ " g-8 \n",
+ " g-9 \n",
+ " g-10 \n",
+ " g-11 \n",
+ " g-12 \n",
+ " g-13 \n",
+ " g-14 \n",
+ " g-15 \n",
+ " g-16 \n",
+ " g-17 \n",
+ " g-18 \n",
+ " g-19 \n",
+ " g-20 \n",
+ " g-21 \n",
+ " g-22 \n",
+ " g-23 \n",
+ " g-24 \n",
+ " g-25 \n",
+ " g-26 \n",
+ " g-27 \n",
+ " g-28 \n",
+ " g-29 \n",
+ " g-30 \n",
+ " g-31 \n",
+ " g-32 \n",
+ " g-33 \n",
+ " g-34 \n",
+ " ... \n",
+ " c-60 \n",
+ " c-61 \n",
+ " c-62 \n",
+ " c-63 \n",
+ " c-64 \n",
+ " c-65 \n",
+ " c-66 \n",
+ " c-67 \n",
+ " c-68 \n",
+ " c-69 \n",
+ " c-70 \n",
+ " c-71 \n",
+ " c-72 \n",
+ " c-73 \n",
+ " c-74 \n",
+ " c-75 \n",
+ " c-76 \n",
+ " c-77 \n",
+ " c-78 \n",
+ " c-79 \n",
+ " c-80 \n",
+ " c-81 \n",
+ " c-82 \n",
+ " c-83 \n",
+ " c-84 \n",
+ " c-85 \n",
+ " c-86 \n",
+ " c-87 \n",
+ " c-88 \n",
+ " c-89 \n",
+ " c-90 \n",
+ " c-91 \n",
+ " c-92 \n",
+ " c-93 \n",
+ " c-94 \n",
+ " c-95 \n",
+ " c-96 \n",
+ " c-97 \n",
+ " c-98 \n",
+ " c-99 \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " id_000644bb2 \n",
+ " com_droga \n",
+ " 24 \n",
+ " D1 \n",
+ " b68db1d53 \n",
+ " 1.0620 \n",
+ " 0.5577 \n",
+ " -0.2479 \n",
+ " -0.6208 \n",
+ " -0.1944 \n",
+ " -1.0120 \n",
+ " -1.0220 \n",
+ " -0.0326 \n",
+ " 0.5548 \n",
+ " -0.0921 \n",
+ " 1.1830 \n",
+ " 0.1530 \n",
+ " 0.5574 \n",
+ " -0.4015 \n",
+ " 0.1789 \n",
+ " -0.6528 \n",
+ " -0.7969 \n",
+ " 0.6342 \n",
+ " 0.1778 \n",
+ " -0.3694 \n",
+ " -0.5688 \n",
+ " -1.1360 \n",
+ " -1.1880 \n",
+ " 0.6940 \n",
+ " 0.4393 \n",
+ " 0.2664 \n",
+ " 0.1907 \n",
+ " 0.1628 \n",
+ " -0.2853 \n",
+ " 0.5819 \n",
+ " 0.2934 \n",
+ " -0.5584 \n",
+ " -0.0916 \n",
+ " -0.3010 \n",
+ " -0.1537 \n",
+ " ... \n",
+ " 0.4805 \n",
+ " 0.4965 \n",
+ " 0.3680 \n",
+ " 0.8427 \n",
+ " 0.1042 \n",
+ " 0.1403 \n",
+ " 0.1758 \n",
+ " 1.2570 \n",
+ " -0.5979 \n",
+ " 1.2250 \n",
+ " -0.0553 \n",
+ " 0.7351 \n",
+ " 0.5810 \n",
+ " 0.9590 \n",
+ " 0.2427 \n",
+ " 0.0495 \n",
+ " 0.4141 \n",
+ " 0.8432 \n",
+ " 0.6162 \n",
+ " -0.7318 \n",
+ " 1.2120 \n",
+ " 0.6362 \n",
+ " -0.4427 \n",
+ " 0.1288 \n",
+ " 1.4840 \n",
+ " 0.1799 \n",
+ " 0.5367 \n",
+ " -0.1111 \n",
+ " -1.0120 \n",
+ " 0.6685 \n",
+ " 0.2862 \n",
+ " 0.2584 \n",
+ " 0.8076 \n",
+ " 0.5523 \n",
+ " -0.1912 \n",
+ " 0.6584 \n",
+ " -0.3981 \n",
+ " 0.2139 \n",
+ " 0.3801 \n",
+ " 0.4176 \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " id_000779bfc \n",
+ " com_droga \n",
+ " 72 \n",
+ " D1 \n",
+ " df89a8e5a \n",
+ " 0.0743 \n",
+ " 0.4087 \n",
+ " 0.2991 \n",
+ " 0.0604 \n",
+ " 1.0190 \n",
+ " 0.5207 \n",
+ " 0.2341 \n",
+ " 0.3372 \n",
+ " -0.4047 \n",
+ " 0.8507 \n",
+ " -1.1520 \n",
+ " -0.4201 \n",
+ " -0.0958 \n",
+ " 0.4590 \n",
+ " 0.0803 \n",
+ " 0.2250 \n",
+ " 0.5293 \n",
+ " 0.2839 \n",
+ " -0.3494 \n",
+ " 0.2883 \n",
+ " 0.9449 \n",
+ " -0.1646 \n",
+ " -0.2657 \n",
+ " -0.3372 \n",
+ " 0.3135 \n",
+ " -0.4316 \n",
+ " 0.4773 \n",
+ " 0.2075 \n",
+ " -0.4216 \n",
+ " -0.1161 \n",
+ " -0.0499 \n",
+ " -0.2627 \n",
+ " 0.9959 \n",
+ " -0.2483 \n",
+ " 0.2655 \n",
+ " ... \n",
+ " 0.4083 \n",
+ " 0.0319 \n",
+ " 0.3905 \n",
+ " 0.7099 \n",
+ " 0.2912 \n",
+ " 0.4151 \n",
+ " -0.2840 \n",
+ " -0.3104 \n",
+ " -0.6373 \n",
+ " 0.2887 \n",
+ " -0.0765 \n",
+ " 0.2539 \n",
+ " 0.4443 \n",
+ " 0.5932 \n",
+ " 0.2031 \n",
+ " 0.7639 \n",
+ " 0.5499 \n",
+ " -0.3322 \n",
+ " -0.0977 \n",
+ " 0.4329 \n",
+ " -0.2782 \n",
+ " 0.7827 \n",
+ " 0.5934 \n",
+ " 0.3402 \n",
+ " 0.1499 \n",
+ " 0.4420 \n",
+ " 0.9366 \n",
+ " 0.8193 \n",
+ " -0.4236 \n",
+ " 0.3192 \n",
+ " -0.4265 \n",
+ " 0.7543 \n",
+ " 0.4708 \n",
+ " 0.0230 \n",
+ " 0.2957 \n",
+ " 0.4899 \n",
+ " 0.1522 \n",
+ " 0.1241 \n",
+ " 0.6077 \n",
+ " 0.7371 \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " id_000a6266a \n",
+ " com_droga \n",
+ " 48 \n",
+ " D1 \n",
+ " 18bb41b2c \n",
+ " 0.6280 \n",
+ " 0.5817 \n",
+ " 1.5540 \n",
+ " -0.0764 \n",
+ " -0.0323 \n",
+ " 1.2390 \n",
+ " 0.1715 \n",
+ " 0.2155 \n",
+ " 0.0065 \n",
+ " 1.2300 \n",
+ " -0.4797 \n",
+ " -0.5631 \n",
+ " -0.0366 \n",
+ " -1.8300 \n",
+ " 0.6057 \n",
+ " -0.3278 \n",
+ " 0.6042 \n",
+ " -0.3075 \n",
+ " -0.1147 \n",
+ " -0.0570 \n",
+ " -0.0799 \n",
+ " -0.8181 \n",
+ " -1.5320 \n",
+ " 0.2307 \n",
+ " 0.4901 \n",
+ " 0.4780 \n",
+ " -1.3970 \n",
+ " 4.6240 \n",
+ " -0.0437 \n",
+ " 1.2870 \n",
+ " -1.8530 \n",
+ " 0.6069 \n",
+ " 0.4290 \n",
+ " 0.1783 \n",
+ " 0.0018 \n",
+ " ... \n",
+ " -0.5477 \n",
+ " -0.7576 \n",
+ " -0.0444 \n",
+ " 0.1894 \n",
+ " -0.0014 \n",
+ " -2.3640 \n",
+ " -0.4682 \n",
+ " 0.1210 \n",
+ " -0.5177 \n",
+ " -0.0604 \n",
+ " 0.1682 \n",
+ " -0.4436 \n",
+ " 0.4963 \n",
+ " 0.1363 \n",
+ " 0.3335 \n",
+ " 0.9760 \n",
+ " -0.0427 \n",
+ " -0.1235 \n",
+ " 0.0959 \n",
+ " 0.0690 \n",
+ " -0.9416 \n",
+ " -0.7548 \n",
+ " -0.1109 \n",
+ " -0.6272 \n",
+ " 0.3019 \n",
+ " 0.1172 \n",
+ " 0.1093 \n",
+ " -0.3113 \n",
+ " 0.3019 \n",
+ " -0.0873 \n",
+ " -0.7250 \n",
+ " -0.6297 \n",
+ " 0.6103 \n",
+ " 0.0223 \n",
+ " -1.3240 \n",
+ " -0.3174 \n",
+ " -0.6417 \n",
+ " -0.2187 \n",
+ " -1.4080 \n",
+ " 0.6931 \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " id_0015fd391 \n",
+ " com_droga \n",
+ " 48 \n",
+ " D1 \n",
+ " 8c7f86626 \n",
+ " -0.5138 \n",
+ " -0.2491 \n",
+ " -0.2656 \n",
+ " 0.5288 \n",
+ " 4.0620 \n",
+ " -0.8095 \n",
+ " -1.9590 \n",
+ " 0.1792 \n",
+ " -0.1321 \n",
+ " -1.0600 \n",
+ " -0.8269 \n",
+ " -0.3584 \n",
+ " -0.8511 \n",
+ " -0.5844 \n",
+ " -2.5690 \n",
+ " 0.8183 \n",
+ " -0.0532 \n",
+ " -0.8554 \n",
+ " 0.1160 \n",
+ " -2.3520 \n",
+ " 2.1200 \n",
+ " -1.1580 \n",
+ " -0.7191 \n",
+ " -0.8004 \n",
+ " -1.4670 \n",
+ " -0.0107 \n",
+ " -0.8995 \n",
+ " 0.2406 \n",
+ " -0.2479 \n",
+ " -1.0890 \n",
+ " -0.7575 \n",
+ " 0.0881 \n",
+ " -2.7370 \n",
+ " 0.8745 \n",
+ " 0.5787 \n",
+ " ... \n",
+ " -2.1220 \n",
+ " -0.3752 \n",
+ " -2.3820 \n",
+ " -3.7350 \n",
+ " -2.9740 \n",
+ " -1.4930 \n",
+ " -1.6600 \n",
+ " -3.1660 \n",
+ " 0.2816 \n",
+ " -0.2990 \n",
+ " -1.1870 \n",
+ " -0.5044 \n",
+ " -1.7750 \n",
+ " -1.6120 \n",
+ " -0.9215 \n",
+ " -1.0810 \n",
+ " -3.0520 \n",
+ " -3.4470 \n",
+ " -2.7740 \n",
+ " -1.8460 \n",
+ " -0.5568 \n",
+ " -3.3960 \n",
+ " -2.9510 \n",
+ " -1.1550 \n",
+ " -3.2620 \n",
+ " -1.5390 \n",
+ " -2.4600 \n",
+ " -0.9417 \n",
+ " -1.5550 \n",
+ " 0.2431 \n",
+ " -2.0990 \n",
+ " -0.6441 \n",
+ " -5.6300 \n",
+ " -1.3780 \n",
+ " -0.8632 \n",
+ " -1.2880 \n",
+ " -1.6210 \n",
+ " -0.8784 \n",
+ " -0.3876 \n",
+ " -0.8154 \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " id_001626bd3 \n",
+ " com_droga \n",
+ " 72 \n",
+ " D2 \n",
+ " 7cbed3131 \n",
+ " -0.3254 \n",
+ " -0.4009 \n",
+ " 0.9700 \n",
+ " 0.6919 \n",
+ " 1.4180 \n",
+ " -0.8244 \n",
+ " -0.2800 \n",
+ " -0.1498 \n",
+ " -0.8789 \n",
+ " 0.8630 \n",
+ " -0.2219 \n",
+ " -0.5121 \n",
+ " -0.9577 \n",
+ " 1.1750 \n",
+ " 0.2042 \n",
+ " 0.1970 \n",
+ " 0.1244 \n",
+ " -1.7090 \n",
+ " -0.3543 \n",
+ " -0.5160 \n",
+ " -0.3330 \n",
+ " -0.2685 \n",
+ " 0.7649 \n",
+ " 0.2057 \n",
+ " 1.3720 \n",
+ " 0.6835 \n",
+ " 0.8056 \n",
+ " -0.3754 \n",
+ " -1.2090 \n",
+ " 0.2965 \n",
+ " -0.0712 \n",
+ " 0.6389 \n",
+ " 0.6674 \n",
+ " -0.0783 \n",
+ " 1.1740 \n",
+ " ... \n",
+ " -0.2274 \n",
+ " 0.3215 \n",
+ " 0.1535 \n",
+ " -0.4640 \n",
+ " -0.5943 \n",
+ " 0.3973 \n",
+ " 0.1500 \n",
+ " 0.5178 \n",
+ " 0.5159 \n",
+ " 0.6091 \n",
+ " 0.1813 \n",
+ " -0.4249 \n",
+ " 0.7832 \n",
+ " 0.6529 \n",
+ " 0.5648 \n",
+ " 0.4817 \n",
+ " 0.0587 \n",
+ " 0.5303 \n",
+ " 0.6376 \n",
+ " -0.3966 \n",
+ " -1.4950 \n",
+ " -0.9625 \n",
+ " -0.0541 \n",
+ " 0.6273 \n",
+ " 0.4563 \n",
+ " 0.0698 \n",
+ " 0.8134 \n",
+ " 0.1924 \n",
+ " 0.6054 \n",
+ " -0.1824 \n",
+ " 0.0042 \n",
+ " 0.0048 \n",
+ " 0.6670 \n",
+ " 1.0690 \n",
+ " 0.5523 \n",
+ " -0.3031 \n",
+ " 0.1094 \n",
+ " 0.2885 \n",
+ " -0.3786 \n",
+ " 0.7125 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
5 rows × 877 columns
\n",
+ "
Para conseguir explorar essa base, precisamos saber quais são as colunas, os tipos de dados que possuem nas colunas, se tem valores nulos ou faltantes e algumas estatísticas"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "ApfYDGeqyLeG",
+ "outputId": "b013108d-4e82-4916-ccbc-e794c16578da"
+ },
+ "source": [
+ "df_experimentos.columns"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "Index(['id', 'tratamento', 'tempo', 'dose', 'droga', 'g-0', 'g-1', 'g-2',\n",
+ " 'g-3', 'g-4',\n",
+ " ...\n",
+ " 'c-90', 'c-91', 'c-92', 'c-93', 'c-94', 'c-95', 'c-96', 'c-97', 'c-98',\n",
+ " 'c-99'],\n",
+ " dtype='object', length=877)"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 146
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "jHIO8a8cyVN3",
+ "outputId": "327e4d1a-62da-4ba5-8f0f-c6188bc39897"
+ },
+ "source": [
+ "#Verificando o tipo das colunas principais\n",
+ "df_experimentos[['id', 'tratamento', 'tempo', 'dose', 'droga','g-0', 'c-0']].info()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "\n",
+ "RangeIndex: 23814 entries, 0 to 23813\n",
+ "Data columns (total 7 columns):\n",
+ " # Column Non-Null Count Dtype \n",
+ "--- ------ -------------- ----- \n",
+ " 0 id 23814 non-null object \n",
+ " 1 tratamento 23814 non-null object \n",
+ " 2 tempo 23814 non-null int64 \n",
+ " 3 dose 23814 non-null object \n",
+ " 4 droga 23814 non-null object \n",
+ " 5 g-0 23814 non-null float64\n",
+ " 6 c-0 23814 non-null float64\n",
+ "dtypes: float64(2), int64(1), object(4)\n",
+ "memory usage: 1.3+ MB\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 333
+ },
+ "id": "PEevXUajObHT",
+ "outputId": "cbe73a06-8e6e-4cde-8995-39a99623d48d"
+ },
+ "source": [
+ "#Verificando a contagem, média, desvio padrão, valor minimo, percentis(25%, 50%, 75%) e valor máximo.\n",
+ "df_experimentos.describe()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " tempo \n",
+ " g-0 \n",
+ " g-1 \n",
+ " g-2 \n",
+ " g-3 \n",
+ " g-4 \n",
+ " g-5 \n",
+ " g-6 \n",
+ " g-7 \n",
+ " g-8 \n",
+ " g-9 \n",
+ " g-10 \n",
+ " g-11 \n",
+ " g-12 \n",
+ " g-13 \n",
+ " g-14 \n",
+ " g-15 \n",
+ " g-16 \n",
+ " g-17 \n",
+ " g-18 \n",
+ " g-19 \n",
+ " g-20 \n",
+ " g-21 \n",
+ " g-22 \n",
+ " g-23 \n",
+ " g-24 \n",
+ " g-25 \n",
+ " g-26 \n",
+ " g-27 \n",
+ " g-28 \n",
+ " g-29 \n",
+ " g-30 \n",
+ " g-31 \n",
+ " g-32 \n",
+ " g-33 \n",
+ " g-34 \n",
+ " g-35 \n",
+ " g-36 \n",
+ " g-37 \n",
+ " g-38 \n",
+ " ... \n",
+ " c-60 \n",
+ " c-61 \n",
+ " c-62 \n",
+ " c-63 \n",
+ " c-64 \n",
+ " c-65 \n",
+ " c-66 \n",
+ " c-67 \n",
+ " c-68 \n",
+ " c-69 \n",
+ " c-70 \n",
+ " c-71 \n",
+ " c-72 \n",
+ " c-73 \n",
+ " c-74 \n",
+ " c-75 \n",
+ " c-76 \n",
+ " c-77 \n",
+ " c-78 \n",
+ " c-79 \n",
+ " c-80 \n",
+ " c-81 \n",
+ " c-82 \n",
+ " c-83 \n",
+ " c-84 \n",
+ " c-85 \n",
+ " c-86 \n",
+ " c-87 \n",
+ " c-88 \n",
+ " c-89 \n",
+ " c-90 \n",
+ " c-91 \n",
+ " c-92 \n",
+ " c-93 \n",
+ " c-94 \n",
+ " c-95 \n",
+ " c-96 \n",
+ " c-97 \n",
+ " c-98 \n",
+ " c-99 \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " count \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " ... \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " \n",
+ " \n",
+ " mean \n",
+ " 48.020156 \n",
+ " 0.248366 \n",
+ " -0.095684 \n",
+ " 0.152253 \n",
+ " 0.081971 \n",
+ " 0.057347 \n",
+ " -0.138836 \n",
+ " 0.035961 \n",
+ " -0.202651 \n",
+ " -0.190083 \n",
+ " 0.119905 \n",
+ " -0.123321 \n",
+ " 0.182307 \n",
+ " 0.143203 \n",
+ " 0.209402 \n",
+ " -0.173884 \n",
+ " -0.024432 \n",
+ " 0.126823 \n",
+ " -0.146663 \n",
+ " 0.087687 \n",
+ " -0.082982 \n",
+ " -0.111908 \n",
+ " -0.087379 \n",
+ " 0.047548 \n",
+ " -0.117474 \n",
+ " -0.113212 \n",
+ " -0.052746 \n",
+ " -0.091055 \n",
+ " 0.112176 \n",
+ " -0.046458 \n",
+ " -0.076239 \n",
+ " -0.197699 \n",
+ " 0.382177 \n",
+ " -0.189432 \n",
+ " 0.078791 \n",
+ " -0.093312 \n",
+ " 0.135729 \n",
+ " -0.188616 \n",
+ " -0.606710 \n",
+ " 0.534425 \n",
+ " ... \n",
+ " -0.517397 \n",
+ " -0.360770 \n",
+ " -0.435752 \n",
+ " -0.613591 \n",
+ " -0.402083 \n",
+ " -0.619682 \n",
+ " -0.452265 \n",
+ " -0.497164 \n",
+ " -0.413836 \n",
+ " -0.277029 \n",
+ " -0.547845 \n",
+ " -0.358611 \n",
+ " -0.442906 \n",
+ " -0.475194 \n",
+ " -0.010404 \n",
+ " -0.467001 \n",
+ " -0.276963 \n",
+ " -0.455848 \n",
+ " -0.412918 \n",
+ " -0.456404 \n",
+ " -0.472514 \n",
+ " -0.505481 \n",
+ " -0.492735 \n",
+ " -0.446836 \n",
+ " -0.463029 \n",
+ " -0.409310 \n",
+ " -0.333124 \n",
+ " -0.295009 \n",
+ " -0.328342 \n",
+ " -0.401615 \n",
+ " -0.469244 \n",
+ " -0.461411 \n",
+ " -0.513256 \n",
+ " -0.500142 \n",
+ " -0.507093 \n",
+ " -0.353726 \n",
+ " -0.463485 \n",
+ " -0.378241 \n",
+ " -0.470252 \n",
+ " -0.301505 \n",
+ " \n",
+ " \n",
+ " std \n",
+ " 19.402807 \n",
+ " 1.393399 \n",
+ " 0.812363 \n",
+ " 1.035731 \n",
+ " 0.950012 \n",
+ " 1.032091 \n",
+ " 1.179388 \n",
+ " 0.882395 \n",
+ " 1.125494 \n",
+ " 1.749885 \n",
+ " 1.087180 \n",
+ " 1.291501 \n",
+ " 1.253604 \n",
+ " 1.234590 \n",
+ " 1.273068 \n",
+ " 1.247178 \n",
+ " 0.659839 \n",
+ " 1.418997 \n",
+ " 1.179688 \n",
+ " 0.743301 \n",
+ " 0.844796 \n",
+ " 1.219529 \n",
+ " 0.824401 \n",
+ " 0.924838 \n",
+ " 0.760159 \n",
+ " 1.203186 \n",
+ " 0.866977 \n",
+ " 1.103765 \n",
+ " 1.001687 \n",
+ " 1.027758 \n",
+ " 1.279399 \n",
+ " 1.302567 \n",
+ " 1.559174 \n",
+ " 0.933514 \n",
+ " 1.172270 \n",
+ " 1.174325 \n",
+ " 1.061719 \n",
+ " 1.397677 \n",
+ " 2.200277 \n",
+ " 2.003317 \n",
+ " ... \n",
+ " 2.122318 \n",
+ " 1.710725 \n",
+ " 1.898871 \n",
+ " 2.307820 \n",
+ " 1.785055 \n",
+ " 2.225596 \n",
+ " 1.991021 \n",
+ " 2.063896 \n",
+ " 1.887001 \n",
+ " 1.459639 \n",
+ " 2.187835 \n",
+ " 1.730634 \n",
+ " 1.924716 \n",
+ " 2.021927 \n",
+ " 1.029820 \n",
+ " 2.004317 \n",
+ " 1.429340 \n",
+ " 1.924263 \n",
+ " 1.888788 \n",
+ " 1.832863 \n",
+ " 2.011396 \n",
+ " 2.091353 \n",
+ " 2.055624 \n",
+ " 1.987476 \n",
+ " 2.014045 \n",
+ " 1.883974 \n",
+ " 1.647241 \n",
+ " 1.634073 \n",
+ " 1.663170 \n",
+ " 1.832794 \n",
+ " 2.000488 \n",
+ " 2.042475 \n",
+ " 2.001714 \n",
+ " 2.107105 \n",
+ " 2.159589 \n",
+ " 1.629291 \n",
+ " 2.059725 \n",
+ " 1.703615 \n",
+ " 1.834828 \n",
+ " 1.407918 \n",
+ " \n",
+ " \n",
+ " min \n",
+ " 24.000000 \n",
+ " -5.513000 \n",
+ " -5.737000 \n",
+ " -9.104000 \n",
+ " -5.998000 \n",
+ " -6.369000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -8.337000 \n",
+ " -10.000000 \n",
+ " -5.870000 \n",
+ " -8.587000 \n",
+ " -5.018000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -4.226000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -5.700000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -8.272000 \n",
+ " -8.184000 \n",
+ " -4.835000 \n",
+ " -7.913000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -2.956000 \n",
+ " -8.356000 \n",
+ " -7.182000 \n",
+ " -10.000000 \n",
+ " -9.261000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " ... \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -9.839000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -6.452000 \n",
+ " -10.000000 \n",
+ " -9.938000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " -10.000000 \n",
+ " \n",
+ " \n",
+ " 25% \n",
+ " 24.000000 \n",
+ " -0.473075 \n",
+ " -0.562200 \n",
+ " -0.437750 \n",
+ " -0.429575 \n",
+ " -0.470925 \n",
+ " -0.602225 \n",
+ " -0.493900 \n",
+ " -0.525175 \n",
+ " -0.511675 \n",
+ " -0.360200 \n",
+ " -0.511475 \n",
+ " -0.489675 \n",
+ " -0.447500 \n",
+ " -0.481200 \n",
+ " -0.607975 \n",
+ " -0.404150 \n",
+ " -0.391950 \n",
+ " -0.513775 \n",
+ " -0.272200 \n",
+ " -0.488675 \n",
+ " -0.524600 \n",
+ " -0.538900 \n",
+ " -0.440375 \n",
+ " -0.508900 \n",
+ " -0.533900 \n",
+ " -0.497700 \n",
+ " -0.512875 \n",
+ " -0.467800 \n",
+ " -0.378300 \n",
+ " -0.505750 \n",
+ " -0.457975 \n",
+ " -0.328200 \n",
+ " -0.600500 \n",
+ " -0.478700 \n",
+ " -0.570525 \n",
+ " -0.481800 \n",
+ " -0.541950 \n",
+ " -0.604100 \n",
+ " -0.470250 \n",
+ " ... \n",
+ " -0.588075 \n",
+ " -0.564025 \n",
+ " -0.561000 \n",
+ " -0.583250 \n",
+ " -0.566500 \n",
+ " -0.603200 \n",
+ " -0.541575 \n",
+ " -0.560825 \n",
+ " -0.555200 \n",
+ " -0.534500 \n",
+ " -0.569100 \n",
+ " -0.558300 \n",
+ " -0.573350 \n",
+ " -0.594275 \n",
+ " -0.389925 \n",
+ " -0.551200 \n",
+ " -0.544150 \n",
+ " -0.575075 \n",
+ " -0.568275 \n",
+ " -0.582650 \n",
+ " -0.558575 \n",
+ " -0.562375 \n",
+ " -0.572800 \n",
+ " -0.561225 \n",
+ " -0.560675 \n",
+ " -0.560100 \n",
+ " -0.533700 \n",
+ " -0.504575 \n",
+ " -0.544275 \n",
+ " -0.569150 \n",
+ " -0.566175 \n",
+ " -0.565975 \n",
+ " -0.589975 \n",
+ " -0.568700 \n",
+ " -0.563775 \n",
+ " -0.567975 \n",
+ " -0.552575 \n",
+ " -0.561000 \n",
+ " -0.592600 \n",
+ " -0.562900 \n",
+ " \n",
+ " \n",
+ " 50% \n",
+ " 48.000000 \n",
+ " -0.008850 \n",
+ " -0.046600 \n",
+ " 0.075200 \n",
+ " 0.008050 \n",
+ " -0.026900 \n",
+ " -0.015650 \n",
+ " -0.000650 \n",
+ " -0.017900 \n",
+ " 0.010000 \n",
+ " 0.160450 \n",
+ " 0.038550 \n",
+ " 0.013800 \n",
+ " 0.060250 \n",
+ " 0.009800 \n",
+ " -0.030100 \n",
+ " 0.000000 \n",
+ " 0.149400 \n",
+ " -0.002200 \n",
+ " 0.000000 \n",
+ " -0.027800 \n",
+ " -0.002800 \n",
+ " -0.069350 \n",
+ " -0.011800 \n",
+ " 0.000000 \n",
+ " 0.018300 \n",
+ " -0.011650 \n",
+ " 0.016100 \n",
+ " 0.037600 \n",
+ " 0.003750 \n",
+ " 0.021750 \n",
+ " 0.027000 \n",
+ " 0.019100 \n",
+ " -0.054000 \n",
+ " 0.023350 \n",
+ " 0.003350 \n",
+ " -0.000600 \n",
+ " 0.015350 \n",
+ " -0.000700 \n",
+ " 0.005800 \n",
+ " ... \n",
+ " -0.017650 \n",
+ " -0.041550 \n",
+ " -0.002950 \n",
+ " -0.012650 \n",
+ " -0.005600 \n",
+ " 0.007650 \n",
+ " 0.004950 \n",
+ " 0.000000 \n",
+ " -0.023800 \n",
+ " -0.011450 \n",
+ " -0.007100 \n",
+ " -0.019500 \n",
+ " -0.019500 \n",
+ " -0.009300 \n",
+ " 0.081550 \n",
+ " -0.006900 \n",
+ " 0.018400 \n",
+ " -0.014650 \n",
+ " -0.014350 \n",
+ " 0.005300 \n",
+ " -0.005300 \n",
+ " -0.004050 \n",
+ " 0.003300 \n",
+ " -0.007900 \n",
+ " -0.004600 \n",
+ " -0.002400 \n",
+ " 0.007850 \n",
+ " -0.005600 \n",
+ " -0.020600 \n",
+ " -0.030000 \n",
+ " -0.009900 \n",
+ " 0.003250 \n",
+ " -0.009100 \n",
+ " -0.013750 \n",
+ " -0.003300 \n",
+ " -0.010250 \n",
+ " -0.001250 \n",
+ " -0.006800 \n",
+ " 0.014000 \n",
+ " -0.019500 \n",
+ " \n",
+ " \n",
+ " 75% \n",
+ " 72.000000 \n",
+ " 0.525700 \n",
+ " 0.403075 \n",
+ " 0.663925 \n",
+ " 0.463400 \n",
+ " 0.465375 \n",
+ " 0.510425 \n",
+ " 0.528725 \n",
+ " 0.411900 \n",
+ " 0.549225 \n",
+ " 0.697775 \n",
+ " 0.525400 \n",
+ " 0.575275 \n",
+ " 0.604450 \n",
+ " 0.575825 \n",
+ " 0.457975 \n",
+ " 0.382475 \n",
+ " 0.829500 \n",
+ " 0.494775 \n",
+ " 0.327800 \n",
+ " 0.400600 \n",
+ " 0.492400 \n",
+ " 0.414875 \n",
+ " 0.433400 \n",
+ " 0.329250 \n",
+ " 0.527700 \n",
+ " 0.461650 \n",
+ " 0.508425 \n",
+ " 0.586450 \n",
+ " 0.431275 \n",
+ " 0.507600 \n",
+ " 0.458075 \n",
+ " 0.471075 \n",
+ " 0.391950 \n",
+ " 0.551300 \n",
+ " 0.503725 \n",
+ " 0.564875 \n",
+ " 0.517025 \n",
+ " 0.460500 \n",
+ " 0.642300 \n",
+ " ... \n",
+ " 0.452675 \n",
+ " 0.427675 \n",
+ " 0.462175 \n",
+ " 0.447975 \n",
+ " 0.447150 \n",
+ " 0.441250 \n",
+ " 0.470600 \n",
+ " 0.458550 \n",
+ " 0.441000 \n",
+ " 0.460075 \n",
+ " 0.460950 \n",
+ " 0.449975 \n",
+ " 0.445200 \n",
+ " 0.473200 \n",
+ " 0.563575 \n",
+ " 0.456350 \n",
+ " 0.493400 \n",
+ " 0.448375 \n",
+ " 0.451975 \n",
+ " 0.463075 \n",
+ " 0.447675 \n",
+ " 0.462000 \n",
+ " 0.468900 \n",
+ " 0.452375 \n",
+ " 0.460475 \n",
+ " 0.461675 \n",
+ " 0.465950 \n",
+ " 0.463400 \n",
+ " 0.450075 \n",
+ " 0.430875 \n",
+ " 0.457750 \n",
+ " 0.461500 \n",
+ " 0.445675 \n",
+ " 0.452900 \n",
+ " 0.470900 \n",
+ " 0.444750 \n",
+ " 0.465225 \n",
+ " 0.446400 \n",
+ " 0.461275 \n",
+ " 0.438650 \n",
+ " \n",
+ " \n",
+ " max \n",
+ " 72.000000 \n",
+ " 10.000000 \n",
+ " 5.039000 \n",
+ " 8.257000 \n",
+ " 10.000000 \n",
+ " 10.000000 \n",
+ " 7.282000 \n",
+ " 7.333000 \n",
+ " 5.473000 \n",
+ " 8.887000 \n",
+ " 6.433000 \n",
+ " 10.000000 \n",
+ " 10.000000 \n",
+ " 10.000000 \n",
+ " 10.000000 \n",
+ " 10.000000 \n",
+ " 4.134000 \n",
+ " 6.418000 \n",
+ " 4.750000 \n",
+ " 8.872000 \n",
+ " 4.081000 \n",
+ " 9.842000 \n",
+ " 5.248000 \n",
+ " 5.942000 \n",
+ " 5.201000 \n",
+ " 10.000000 \n",
+ " 8.494000 \n",
+ " 10.000000 \n",
+ " 10.000000 \n",
+ " 10.000000 \n",
+ " 9.416000 \n",
+ " 10.000000 \n",
+ " 10.000000 \n",
+ " 6.796000 \n",
+ " 10.000000 \n",
+ " 10.000000 \n",
+ " 10.000000 \n",
+ " 5.834000 \n",
+ " 5.602000 \n",
+ " 10.000000 \n",
+ " ... \n",
+ " 3.888000 \n",
+ " 3.596000 \n",
+ " 4.857000 \n",
+ " 3.549000 \n",
+ " 3.382000 \n",
+ " 3.328000 \n",
+ " 4.157000 \n",
+ " 3.736000 \n",
+ " 3.582000 \n",
+ " 3.119000 \n",
+ " 3.323000 \n",
+ " 5.014000 \n",
+ " 2.898000 \n",
+ " 4.185000 \n",
+ " 3.170000 \n",
+ " 3.276000 \n",
+ " 4.992000 \n",
+ " 3.770000 \n",
+ " 2.851000 \n",
+ " 3.211000 \n",
+ " 4.534000 \n",
+ " 3.890000 \n",
+ " 3.994000 \n",
+ " 4.321000 \n",
+ " 4.020000 \n",
+ " 3.738000 \n",
+ " 3.252000 \n",
+ " 5.406000 \n",
+ " 3.110000 \n",
+ " 3.320000 \n",
+ " 4.069000 \n",
+ " 3.960000 \n",
+ " 3.927000 \n",
+ " 3.596000 \n",
+ " 3.747000 \n",
+ " 2.814000 \n",
+ " 3.505000 \n",
+ " 2.924000 \n",
+ " 3.111000 \n",
+ " 3.805000 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
8 rows × 873 columns
\n",
+ "
Vamos verificar se a base possui algum valor nulo. Para isso vamos utilizar o `.isnull()` depois somar todos os nulos com a `sum()` e ordenalos de forma decrescente para ver os maiores primeiro com `sort_values()`."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "igFtbtOKOhUW",
+ "outputId": "05e3d2d5-c369-42bb-8edb-9b782f5a7743"
+ },
+ "source": [
+ "df_experimentos.isnull().sum().sort_values(ascending = False)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "c-99 0\n",
+ "g-282 0\n",
+ "g-293 0\n",
+ "g-292 0\n",
+ "g-291 0\n",
+ " ..\n",
+ "g-575 0\n",
+ "g-574 0\n",
+ "g-573 0\n",
+ "g-572 0\n",
+ "id 0\n",
+ "Length: 877, dtype: int64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 149
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "5P8czA9vSsHA"
+ },
+ "source": [
+ "Vamos que não possui nenhum valor nulo na base `Experimentos`"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "8jrB_9DAyXRa",
+ "outputId": "5587760b-feec-4937-ee2a-caa696c246f0"
+ },
+ "source": [
+ "#verificando os possíveis valores e a quantidade da coluna tratamento\n",
+ "df_experimentos['tratamento'].value_counts()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "com_droga 21948\n",
+ "com_controle 1866\n",
+ "Name: tratamento, dtype: int64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 150
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "nxMIwdmnyaSb"
+ },
+ "source": [
+ "Vamos fazer o nosso primeiro Gráfico, vamos verificar a coluna `'Tratamento'`, será que essa coluna está balanceada?"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 380
+ },
+ "id": "uoIAoxH7yZnD",
+ "outputId": "579df7bc-a76b-4d52-d85b-9b40ff01f1b1"
+ },
+ "source": [
+ "sns.countplot(x='tratamento', data = df_experimentos, palette=\"crest\")\n",
+ "plt.title('Tipos de Tratamentos', fontdict={'fontsize':18, 'fontweight': 'bold'},)\n",
+ "plt.xlabel('Tratamento') \n",
+ "plt.ylabel('Quantidade') \n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "B81l5aQzy4pS"
+ },
+ "source": [
+ "Conforme o gráfico vimos que essa coluna está desbalanceada, será que isso afetará na nossa maquina preditiva?\n",
+ "\n",
+ "vamos ver a porcentagem de cada variável dessa coluna, para isso utilizaremos o `value_counts` com o parametro `Normalize`"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "l0Skr-uIz8my",
+ "outputId": "a8f5ac14-1ce2-48b2-9866-3b2adb24842e"
+ },
+ "source": [
+ "print(df_experimentos['tratamento'].value_counts(normalize = True)*100)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "com_droga 92.164273\n",
+ "com_controle 7.835727\n",
+ "Name: tratamento, dtype: float64\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "4KNOi9v_J5n_"
+ },
+ "source": [
+ "Temos uma frequencia total de **92%** de `com_droga` e **7%** de `com_controle`."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "RR2hDZ5P0CnU"
+ },
+ "source": [
+ "Agora vamos ver os gráficos das colunas: Dose e Tempo\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 380
+ },
+ "id": "K7U1M1hi0VKu",
+ "outputId": "f71fa0ed-48d5-47f1-b8a9-04884508f76e"
+ },
+ "source": [
+ "sns.countplot(x = 'dose', data=df_experimentos, palette=\"crest\")\n",
+ "plt.title('Tipos de doses',fontdict={'fontsize':18, 'fontweight': 'bold'},)\n",
+ "plt.xlabel('Doses') \n",
+ "plt.ylabel('Quantidade') \n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 380
+ },
+ "id": "8q7Txm7C0O85",
+ "outputId": "f4723929-e07a-40b6-a955-89490f9093a3"
+ },
+ "source": [
+ "sns.countplot(x='tempo', data = df_experimentos, palette=\"crest\")\n",
+ "plt.title('Tipos utilizados no conjunto de dados', fontdict={'fontsize':18, 'fontweight': 'bold'},)\n",
+ "plt.xlabel('tempo') \n",
+ "plt.ylabel('Quantidade') \n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "49CpmZbj0OZT"
+ },
+ "source": [
+ "Vimos que diferente da coluna `Tratamento`, as colunas `Dose` e `Tempo` estão mais balanceados.\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "iFa-6HwLK-rO"
+ },
+ "source": [
+ "Vamos ver o top 5 drogas mais utilizadas nesse dataset"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "C0MSWtpxK2By",
+ "outputId": "bb8a3d09-1581-42aa-db34-1a5548f0b426"
+ },
+ "source": [
+ "ranking_drogas = df_experimentos['droga'].value_counts().index[0:5]\n",
+ "ranking_drogas"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "Index(['cacb2b860', '87d714366', '9f80f3f77', '8b87a7a83', '5628cb3ee'], dtype='object')"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 155
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 407
+ },
+ "id": "E452ViqdKrsG",
+ "outputId": "b037e287-6be3-495f-e899-2adaa9eb2ad6"
+ },
+ "source": [
+ "sns.color_palette(\"crest\", as_cmap=True)\n",
+ "plt.figure(figsize=(8, 6))\n",
+ "ax = sns.countplot(x = 'droga', data=df_experimentos.query('droga in @ranking_drogas'), color='#092A32')\n",
+ "ax.set_title('Drogas mais utilizadas',fontdict={'fontsize':18, 'fontweight': 'bold'})\n",
+ "plt.xlabel('Drogas') \n",
+ "plt.ylabel('Quantidade') \n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "hkROloRG076Q"
+ },
+ "source": [
+ "Visando deixar mais fácil, vou remover dos nomes das colunas o `-`, por exemplo: 'g-0', após a mudança ficará 'g0'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "TSYoGAon1UXE"
+ },
+ "source": [
+ "df_experimentos.columns = df_experimentos.columns.str.replace(\"-\", \"\",)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "rGjBCv6PLsC-"
+ },
+ "source": [
+ "###Correlação "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "p7tyl_2tMmv_"
+ },
+ "source": [
+ "\"O coeficiente de correlação de Pearson (r) varia entre -1 e +1, cujos valores próximos de -1 e +1 indicam forte correlação linear e próximos de 0 indicam ausência de correlação linear. Note que ele capta apenas relações lineares entre variáveis (quaisquer outras relações, tal coeficiente não é indicado: isso será exemplificado na sequência).\n",
+ "\n",
+ "Note que entre 0 e 1, existe uma grande gama de valores que o coeficiente pode assumir. Para tal, diferentes autores buscaram dar “nomes” aos diferentes valores que o coeficiente de correlação pode assumir, para poder dizer se um dado valor de correlação pode ser dito como fraco/moderada/forte.\"\n",
+ "\n",
+ "Fonte: https://gpestatistica.netlify.app/blog/correlacao/"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "tlHVpFA0LZ9I"
+ },
+ "source": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "Zmpr2DANKfgl"
+ },
+ "source": [
+ "corr_c = df_experimentos.loc[:,'c49':'c99'].corr()"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "oWPoSM9q1wdr",
+ "outputId": "fe7e618f-97bb-43a7-94cf-a00d1653a0c0"
+ },
+ "source": [
+ "# Generate a mask for the upper triangle\n",
+ "mask = np.triu(np.ones_like(corr_c, dtype=bool))\n",
+ "\n",
+ "# Set up the matplotlib figure\n",
+ "f, ax = plt.subplots(figsize=(11, 9))\n",
+ "\n",
+ "# Generate a custom diverging colormap\n",
+ "cmap = sns.color_palette(\"crest\", as_cmap=True)\n",
+ "\n",
+ "# Draw the heatmap with the mask and correct aspect ratio\n",
+ "sns.heatmap(corr_c, mask=mask, cmap=cmap, center=0,\n",
+ " square=True, linewidths=.5, cbar_kws={\"shrink\": .5})\n",
+ "plt.title('Correlações entre c49 até c99',fontdict={'fontsize':18, 'fontweight': 'bold'})\n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "2kzVUPeF2NrG"
+ },
+ "source": [
+ "##Base Resultados "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "DsCTrRSI2UQD"
+ },
+ "source": [
+ "Como já colocamos no `paredão` a ***base de experimento*** e exploramos, vamos colocar no `paredão` a ***base de resultados*** para explorar."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "dPb5EVv-3auK"
+ },
+ "source": [
+ "Como fizemos na exploração anterior vamos ver se essa base tem a mesma **dimensão** que a base de experimento, para isso vamos utilizar o `.shape`"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "pORhjwKK2SuK",
+ "outputId": "bebedbac-893f-476c-9574-faa81e0f1478"
+ },
+ "source": [
+ "df_resultados.shape"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "(23814, 207)"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 162
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "8oh2U1jB5lBc"
+ },
+ "source": [
+ "Diferente da base de dados anterior, essa base tem menos colunas, um total de `207 colunas`. Vamos verificar as primeira linhas."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 261
+ },
+ "id": "ucHK3IoW2C5b",
+ "outputId": "f37f19aa-ea0f-4898-bb99-fc94d5b2be11"
+ },
+ "source": [
+ "df_resultados.head()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " id \n",
+ " 5-alpha_reductase_inhibitor \n",
+ " 11-beta-hsd1_inhibitor \n",
+ " acat_inhibitor \n",
+ " acetylcholine_receptor_agonist \n",
+ " acetylcholine_receptor_antagonist \n",
+ " acetylcholinesterase_inhibitor \n",
+ " adenosine_receptor_agonist \n",
+ " adenosine_receptor_antagonist \n",
+ " adenylyl_cyclase_activator \n",
+ " adrenergic_receptor_agonist \n",
+ " adrenergic_receptor_antagonist \n",
+ " akt_inhibitor \n",
+ " aldehyde_dehydrogenase_inhibitor \n",
+ " alk_inhibitor \n",
+ " ampk_activator \n",
+ " analgesic \n",
+ " androgen_receptor_agonist \n",
+ " androgen_receptor_antagonist \n",
+ " anesthetic_-_local \n",
+ " angiogenesis_inhibitor \n",
+ " angiotensin_receptor_antagonist \n",
+ " anti-inflammatory \n",
+ " antiarrhythmic \n",
+ " antibiotic \n",
+ " anticonvulsant \n",
+ " antifungal \n",
+ " antihistamine \n",
+ " antimalarial \n",
+ " antioxidant \n",
+ " antiprotozoal \n",
+ " antiviral \n",
+ " apoptosis_stimulant \n",
+ " aromatase_inhibitor \n",
+ " atm_kinase_inhibitor \n",
+ " atp-sensitive_potassium_channel_antagonist \n",
+ " atp_synthase_inhibitor \n",
+ " atpase_inhibitor \n",
+ " atr_kinase_inhibitor \n",
+ " aurora_kinase_inhibitor \n",
+ " ... \n",
+ " protein_synthesis_inhibitor \n",
+ " protein_tyrosine_kinase_inhibitor \n",
+ " radiopaque_medium \n",
+ " raf_inhibitor \n",
+ " ras_gtpase_inhibitor \n",
+ " retinoid_receptor_agonist \n",
+ " retinoid_receptor_antagonist \n",
+ " rho_associated_kinase_inhibitor \n",
+ " ribonucleoside_reductase_inhibitor \n",
+ " rna_polymerase_inhibitor \n",
+ " serotonin_receptor_agonist \n",
+ " serotonin_receptor_antagonist \n",
+ " serotonin_reuptake_inhibitor \n",
+ " sigma_receptor_agonist \n",
+ " sigma_receptor_antagonist \n",
+ " smoothened_receptor_antagonist \n",
+ " sodium_channel_inhibitor \n",
+ " sphingosine_receptor_agonist \n",
+ " src_inhibitor \n",
+ " steroid \n",
+ " syk_inhibitor \n",
+ " tachykinin_antagonist \n",
+ " tgf-beta_receptor_inhibitor \n",
+ " thrombin_inhibitor \n",
+ " thymidylate_synthase_inhibitor \n",
+ " tlr_agonist \n",
+ " tlr_antagonist \n",
+ " tnf_inhibitor \n",
+ " topoisomerase_inhibitor \n",
+ " transient_receptor_potential_channel_antagonist \n",
+ " tropomyosin_receptor_kinase_inhibitor \n",
+ " trpv_agonist \n",
+ " trpv_antagonist \n",
+ " tubulin_inhibitor \n",
+ " tyrosine_kinase_inhibitor \n",
+ " ubiquitin_specific_protease_inhibitor \n",
+ " vegfr_inhibitor \n",
+ " vitamin_b \n",
+ " vitamin_d_receptor_agonist \n",
+ " wnt_inhibitor \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " id_000644bb2 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " id_000779bfc \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " id_000a6266a \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " id_0015fd391 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " id_001626bd3 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
5 rows × 207 columns
\n",
+ "
\n",
+ "RangeIndex: 23814 entries, 0 to 23813\n",
+ "Columns: 207 entries, id to wnt_inhibitor\n",
+ "dtypes: int64(206), object(1)\n",
+ "memory usage: 37.6+ MB\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "jJo4VSSr6gc3"
+ },
+ "source": [
+ "No .head() vimos que aparecem muitos 0 e 1, pois 0 significa que não houve um ativamento do mecanismo de Ação, já o 1 é o ativimente do MoA. Vale ressaltar, que o mesmo experimento pode ativar nenhum, um ou vários MoAs."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 350
+ },
+ "id": "Ih1mdvTYTo2G",
+ "outputId": "51170f43-d048-4469-d4c0-3a7aa72c3473"
+ },
+ "source": [
+ "#Verificando a contagem, média, desvio padrão, valor minimo, percentis(25%, 50%, 75%) e valor máximo.\n",
+ "df_resultados.describe()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " 5-alpha_reductase_inhibitor \n",
+ " 11-beta-hsd1_inhibitor \n",
+ " acat_inhibitor \n",
+ " acetylcholine_receptor_agonist \n",
+ " acetylcholine_receptor_antagonist \n",
+ " acetylcholinesterase_inhibitor \n",
+ " adenosine_receptor_agonist \n",
+ " adenosine_receptor_antagonist \n",
+ " adenylyl_cyclase_activator \n",
+ " adrenergic_receptor_agonist \n",
+ " adrenergic_receptor_antagonist \n",
+ " akt_inhibitor \n",
+ " aldehyde_dehydrogenase_inhibitor \n",
+ " alk_inhibitor \n",
+ " ampk_activator \n",
+ " analgesic \n",
+ " androgen_receptor_agonist \n",
+ " androgen_receptor_antagonist \n",
+ " anesthetic_-_local \n",
+ " angiogenesis_inhibitor \n",
+ " angiotensin_receptor_antagonist \n",
+ " anti-inflammatory \n",
+ " antiarrhythmic \n",
+ " antibiotic \n",
+ " anticonvulsant \n",
+ " antifungal \n",
+ " antihistamine \n",
+ " antimalarial \n",
+ " antioxidant \n",
+ " antiprotozoal \n",
+ " antiviral \n",
+ " apoptosis_stimulant \n",
+ " aromatase_inhibitor \n",
+ " atm_kinase_inhibitor \n",
+ " atp-sensitive_potassium_channel_antagonist \n",
+ " atp_synthase_inhibitor \n",
+ " atpase_inhibitor \n",
+ " atr_kinase_inhibitor \n",
+ " aurora_kinase_inhibitor \n",
+ " autotaxin_inhibitor \n",
+ " ... \n",
+ " protein_synthesis_inhibitor \n",
+ " protein_tyrosine_kinase_inhibitor \n",
+ " radiopaque_medium \n",
+ " raf_inhibitor \n",
+ " ras_gtpase_inhibitor \n",
+ " retinoid_receptor_agonist \n",
+ " retinoid_receptor_antagonist \n",
+ " rho_associated_kinase_inhibitor \n",
+ " ribonucleoside_reductase_inhibitor \n",
+ " rna_polymerase_inhibitor \n",
+ " serotonin_receptor_agonist \n",
+ " serotonin_receptor_antagonist \n",
+ " serotonin_reuptake_inhibitor \n",
+ " sigma_receptor_agonist \n",
+ " sigma_receptor_antagonist \n",
+ " smoothened_receptor_antagonist \n",
+ " sodium_channel_inhibitor \n",
+ " sphingosine_receptor_agonist \n",
+ " src_inhibitor \n",
+ " steroid \n",
+ " syk_inhibitor \n",
+ " tachykinin_antagonist \n",
+ " tgf-beta_receptor_inhibitor \n",
+ " thrombin_inhibitor \n",
+ " thymidylate_synthase_inhibitor \n",
+ " tlr_agonist \n",
+ " tlr_antagonist \n",
+ " tnf_inhibitor \n",
+ " topoisomerase_inhibitor \n",
+ " transient_receptor_potential_channel_antagonist \n",
+ " tropomyosin_receptor_kinase_inhibitor \n",
+ " trpv_agonist \n",
+ " trpv_antagonist \n",
+ " tubulin_inhibitor \n",
+ " tyrosine_kinase_inhibitor \n",
+ " ubiquitin_specific_protease_inhibitor \n",
+ " vegfr_inhibitor \n",
+ " vitamin_b \n",
+ " vitamin_d_receptor_agonist \n",
+ " wnt_inhibitor \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " count \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " ... \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.00000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " 23814.000000 \n",
+ " \n",
+ " \n",
+ " mean \n",
+ " 0.000714 \n",
+ " 0.000756 \n",
+ " 0.001008 \n",
+ " 0.007979 \n",
+ " 0.012640 \n",
+ " 0.003065 \n",
+ " 0.002268 \n",
+ " 0.004031 \n",
+ " 0.000504 \n",
+ " 0.011338 \n",
+ " 0.015117 \n",
+ " 0.002771 \n",
+ " 0.000294 \n",
+ " 0.001764 \n",
+ " 0.000504 \n",
+ " 0.000504 \n",
+ " 0.002016 \n",
+ " 0.003737 \n",
+ " 0.003359 \n",
+ " 0.001512 \n",
+ " 0.001554 \n",
+ " 0.003065 \n",
+ " 0.000252 \n",
+ " 0.001806 \n",
+ " 0.000504 \n",
+ " 0.000546 \n",
+ " 0.000504 \n",
+ " 0.000756 \n",
+ " 0.003065 \n",
+ " 0.001512 \n",
+ " 0.000966 \n",
+ " 0.002058 \n",
+ " 0.001974 \n",
+ " 0.000252 \n",
+ " 0.000042 \n",
+ " 0.000504 \n",
+ " 0.004073 \n",
+ " 0.000798 \n",
+ " 0.004031 \n",
+ " 0.000252 \n",
+ " ... \n",
+ " 0.004325 \n",
+ " 0.000798 \n",
+ " 0.002352 \n",
+ " 0.009364 \n",
+ " 0.000504 \n",
+ " 0.002813 \n",
+ " 0.000252 \n",
+ " 0.00147 \n",
+ " 0.001554 \n",
+ " 0.001050 \n",
+ " 0.009910 \n",
+ " 0.016965 \n",
+ " 0.001848 \n",
+ " 0.001512 \n",
+ " 0.001512 \n",
+ " 0.001050 \n",
+ " 0.011212 \n",
+ " 0.001050 \n",
+ " 0.002981 \n",
+ " 0.000252 \n",
+ " 0.000798 \n",
+ " 0.002520 \n",
+ " 0.001260 \n",
+ " 0.000798 \n",
+ " 0.001554 \n",
+ " 0.001260 \n",
+ " 0.000294 \n",
+ " 0.001512 \n",
+ " 0.005333 \n",
+ " 0.000756 \n",
+ " 0.000252 \n",
+ " 0.001050 \n",
+ " 0.002016 \n",
+ " 0.013270 \n",
+ " 0.003065 \n",
+ " 0.000252 \n",
+ " 0.007139 \n",
+ " 0.001092 \n",
+ " 0.001638 \n",
+ " 0.001260 \n",
+ " \n",
+ " \n",
+ " std \n",
+ " 0.026709 \n",
+ " 0.027483 \n",
+ " 0.031731 \n",
+ " 0.088967 \n",
+ " 0.111716 \n",
+ " 0.055283 \n",
+ " 0.047566 \n",
+ " 0.063365 \n",
+ " 0.022443 \n",
+ " 0.105876 \n",
+ " 0.122022 \n",
+ " 0.052573 \n",
+ " 0.017143 \n",
+ " 0.041960 \n",
+ " 0.022443 \n",
+ " 0.022443 \n",
+ " 0.044851 \n",
+ " 0.061020 \n",
+ " 0.057864 \n",
+ " 0.038852 \n",
+ " 0.039387 \n",
+ " 0.055283 \n",
+ " 0.015871 \n",
+ " 0.042456 \n",
+ " 0.022443 \n",
+ " 0.023359 \n",
+ " 0.022443 \n",
+ " 0.027483 \n",
+ " 0.055283 \n",
+ " 0.038852 \n",
+ " 0.031063 \n",
+ " 0.045315 \n",
+ " 0.044383 \n",
+ " 0.015871 \n",
+ " 0.006480 \n",
+ " 0.022443 \n",
+ " 0.063693 \n",
+ " 0.028236 \n",
+ " 0.063365 \n",
+ " 0.015871 \n",
+ " ... \n",
+ " 0.065625 \n",
+ " 0.028236 \n",
+ " 0.048437 \n",
+ " 0.096317 \n",
+ " 0.022443 \n",
+ " 0.052969 \n",
+ " 0.015871 \n",
+ " 0.03831 \n",
+ " 0.039387 \n",
+ " 0.032384 \n",
+ " 0.099057 \n",
+ " 0.129142 \n",
+ " 0.042946 \n",
+ " 0.038852 \n",
+ " 0.038852 \n",
+ " 0.032384 \n",
+ " 0.105293 \n",
+ " 0.032384 \n",
+ " 0.054522 \n",
+ " 0.015871 \n",
+ " 0.028236 \n",
+ " 0.050133 \n",
+ " 0.035472 \n",
+ " 0.028236 \n",
+ " 0.039387 \n",
+ " 0.035472 \n",
+ " 0.017143 \n",
+ " 0.038852 \n",
+ " 0.072834 \n",
+ " 0.027483 \n",
+ " 0.015871 \n",
+ " 0.032384 \n",
+ " 0.044851 \n",
+ " 0.114429 \n",
+ " 0.055283 \n",
+ " 0.015871 \n",
+ " 0.084190 \n",
+ " 0.033025 \n",
+ " 0.040436 \n",
+ " 0.035472 \n",
+ " \n",
+ " \n",
+ " min \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " ... \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.00000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " \n",
+ " \n",
+ " 25% \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " ... \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.00000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " \n",
+ " \n",
+ " 50% \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " ... \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.00000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " \n",
+ " \n",
+ " 75% \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " ... \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.00000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " 0.000000 \n",
+ " \n",
+ " \n",
+ " max \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " ... \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.00000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " 1.000000 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
8 rows × 206 columns
\n",
+ "
\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " moa \n",
+ " quantidade \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " nfkb_inhibitor \n",
+ " 832 \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " proteasome_inhibitor \n",
+ " 726 \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " cyclooxygenase_inhibitor \n",
+ " 435 \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " dopamine_receptor_antagonist \n",
+ " 424 \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " serotonin_receptor_antagonist \n",
+ " 404 \n",
+ " \n",
+ " \n",
+ " ... \n",
+ " ... \n",
+ " ... \n",
+ " \n",
+ " \n",
+ " 201 \n",
+ " protein_phosphatase_inhibitor \n",
+ " 6 \n",
+ " \n",
+ " \n",
+ " 202 \n",
+ " autotaxin_inhibitor \n",
+ " 6 \n",
+ " \n",
+ " \n",
+ " 203 \n",
+ " diuretic \n",
+ " 6 \n",
+ " \n",
+ " \n",
+ " 204 \n",
+ " erbb2_inhibitor \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 205 \n",
+ " atp-sensitive_potassium_channel_antagonist \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
206 rows × 2 columns
\n",
+ "
\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " moa \n",
+ " quantidade \n",
+ " action \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " nfkb_inhibitor \n",
+ " 832 \n",
+ " inhibitor \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " proteasome_inhibitor \n",
+ " 726 \n",
+ " inhibitor \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " cyclooxygenase_inhibitor \n",
+ " 435 \n",
+ " inhibitor \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " dopamine_receptor_antagonist \n",
+ " 424 \n",
+ " antagonist \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " serotonin_receptor_antagonist \n",
+ " 404 \n",
+ " antagonist \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " quantidade \n",
+ " \n",
+ " \n",
+ " action \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " inhibitor \n",
+ " 9693 \n",
+ " \n",
+ " \n",
+ " antagonist \n",
+ " 3449 \n",
+ " \n",
+ " \n",
+ " agonist \n",
+ " 2330 \n",
+ " \n",
+ " \n",
+ " blocker \n",
+ " 323 \n",
+ " \n",
+ " \n",
+ " agent \n",
+ " 150 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ },
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "IDa45JA8po48"
+ },
+ "source": [
+ "#***Pré-processamento dos Dados*** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "1CfURjV8FKMF"
+ },
+ "source": [
+ "Agora que já colocamos as duas bases no `paredawn`, vamos preparar nossos dados para a criação da máquina preditiva"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Ipl6TKtXp0ZB"
+ },
+ "source": [
+ "##Transformação das colunas "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "npQBCR0nIdRh"
+ },
+ "source": [
+ "Vamos transformar as colunas que possuem o tipo de dados string em `1-0`\n",
+ "\n",
+ "Para a Base Experimentos vamos transformar as colunas: `tempo`, `tratamento`, `dose` e criar as colunas para armazenar essas transformaçoes em `tempo_24` , `tempo_48`, `tempo_72`, `dose`."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "aOWAzIDqrob8"
+ },
+ "source": [
+ "###Base Experimentos"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Cu1W6YFuFjfa"
+ },
+ "source": [
+ "Vamos separar a coluna tempo em 3 `tempo_24`, `tempo_48`, `tempo_72`, cada coluna será uma coluna binária (`0-1`).\n",
+ "\n",
+ "Caso seja o tempo que esteja no nome da coluna, será identificado como `1`, caso não é `0`"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "Od1Tk3sgtBCR",
+ "outputId": "125b54ed-7649-4af7-ab83-29ffe67d8a59"
+ },
+ "source": [
+ "tempo24 = df_experimentos['tempo'] == 24\n",
+ "tempo24"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "0 True\n",
+ "1 False\n",
+ "2 False\n",
+ "3 False\n",
+ "4 False\n",
+ " ... \n",
+ "23809 True\n",
+ "23810 True\n",
+ "23811 False\n",
+ "23812 True\n",
+ "23813 False\n",
+ "Name: tempo, Length: 23814, dtype: bool"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 175
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "pKXKf2gHtTMI",
+ "outputId": "3c0d41f7-c45b-4f67-b984-23fa1567ac0c"
+ },
+ "source": [
+ "tempo48 = df_experimentos['tempo'] == 48\n",
+ "tempo48"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "0 False\n",
+ "1 False\n",
+ "2 True\n",
+ "3 True\n",
+ "4 False\n",
+ " ... \n",
+ "23809 False\n",
+ "23810 False\n",
+ "23811 True\n",
+ "23812 False\n",
+ "23813 False\n",
+ "Name: tempo, Length: 23814, dtype: bool"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 176
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "Aib5Gs_ktXMc"
+ },
+ "source": [
+ "tempo72 = df_experimentos['tempo'] == 72"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "QW6yus0ntafs"
+ },
+ "source": [
+ "com_droga = df_experimentos['tratamento'] =='com_droga'"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "dfYMlqHCtiDs"
+ },
+ "source": [
+ "D1 = df_experimentos['dose'] == 'D1'"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "GWYnF72BtZiQ"
+ },
+ "source": [
+ "df_experimentos['tempo_24'] = tempo24.astype('int')\n",
+ "df_experimentos['tempo_48'] = tempo48.astype('int')\n",
+ "df_experimentos['tempo_72'] = tempo72.astype('int')\n",
+ "df_experimentos['com_droga'] = com_droga.astype('int')\n",
+ "df_experimentos['dose'] = D1.astype('int')"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "Vqed45VLtldV",
+ "outputId": "1d55bc76-57f0-4514-a8a7-3e558b68fb9a"
+ },
+ "source": [
+ "df_experimentos.head()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " id \n",
+ " tratamento \n",
+ " tempo \n",
+ " dose \n",
+ " droga \n",
+ " g0 \n",
+ " g1 \n",
+ " g2 \n",
+ " g3 \n",
+ " g4 \n",
+ " g5 \n",
+ " g6 \n",
+ " g7 \n",
+ " g8 \n",
+ " g9 \n",
+ " g10 \n",
+ " g11 \n",
+ " g12 \n",
+ " g13 \n",
+ " g14 \n",
+ " g15 \n",
+ " g16 \n",
+ " g17 \n",
+ " g18 \n",
+ " g19 \n",
+ " g20 \n",
+ " g21 \n",
+ " g22 \n",
+ " g23 \n",
+ " g24 \n",
+ " g25 \n",
+ " g26 \n",
+ " g27 \n",
+ " g28 \n",
+ " g29 \n",
+ " g30 \n",
+ " g31 \n",
+ " g32 \n",
+ " g33 \n",
+ " g34 \n",
+ " ... \n",
+ " c64 \n",
+ " c65 \n",
+ " c66 \n",
+ " c67 \n",
+ " c68 \n",
+ " c69 \n",
+ " c70 \n",
+ " c71 \n",
+ " c72 \n",
+ " c73 \n",
+ " c74 \n",
+ " c75 \n",
+ " c76 \n",
+ " c77 \n",
+ " c78 \n",
+ " c79 \n",
+ " c80 \n",
+ " c81 \n",
+ " c82 \n",
+ " c83 \n",
+ " c84 \n",
+ " c85 \n",
+ " c86 \n",
+ " c87 \n",
+ " c88 \n",
+ " c89 \n",
+ " c90 \n",
+ " c91 \n",
+ " c92 \n",
+ " c93 \n",
+ " c94 \n",
+ " c95 \n",
+ " c96 \n",
+ " c97 \n",
+ " c98 \n",
+ " c99 \n",
+ " tempo_24 \n",
+ " tempo_48 \n",
+ " tempo_72 \n",
+ " com_droga \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " id_000644bb2 \n",
+ " com_droga \n",
+ " 24 \n",
+ " 1 \n",
+ " b68db1d53 \n",
+ " 1.0620 \n",
+ " 0.5577 \n",
+ " -0.2479 \n",
+ " -0.6208 \n",
+ " -0.1944 \n",
+ " -1.0120 \n",
+ " -1.0220 \n",
+ " -0.0326 \n",
+ " 0.5548 \n",
+ " -0.0921 \n",
+ " 1.1830 \n",
+ " 0.1530 \n",
+ " 0.5574 \n",
+ " -0.4015 \n",
+ " 0.1789 \n",
+ " -0.6528 \n",
+ " -0.7969 \n",
+ " 0.6342 \n",
+ " 0.1778 \n",
+ " -0.3694 \n",
+ " -0.5688 \n",
+ " -1.1360 \n",
+ " -1.1880 \n",
+ " 0.6940 \n",
+ " 0.4393 \n",
+ " 0.2664 \n",
+ " 0.1907 \n",
+ " 0.1628 \n",
+ " -0.2853 \n",
+ " 0.5819 \n",
+ " 0.2934 \n",
+ " -0.5584 \n",
+ " -0.0916 \n",
+ " -0.3010 \n",
+ " -0.1537 \n",
+ " ... \n",
+ " 0.1042 \n",
+ " 0.1403 \n",
+ " 0.1758 \n",
+ " 1.2570 \n",
+ " -0.5979 \n",
+ " 1.2250 \n",
+ " -0.0553 \n",
+ " 0.7351 \n",
+ " 0.5810 \n",
+ " 0.9590 \n",
+ " 0.2427 \n",
+ " 0.0495 \n",
+ " 0.4141 \n",
+ " 0.8432 \n",
+ " 0.6162 \n",
+ " -0.7318 \n",
+ " 1.2120 \n",
+ " 0.6362 \n",
+ " -0.4427 \n",
+ " 0.1288 \n",
+ " 1.4840 \n",
+ " 0.1799 \n",
+ " 0.5367 \n",
+ " -0.1111 \n",
+ " -1.0120 \n",
+ " 0.6685 \n",
+ " 0.2862 \n",
+ " 0.2584 \n",
+ " 0.8076 \n",
+ " 0.5523 \n",
+ " -0.1912 \n",
+ " 0.6584 \n",
+ " -0.3981 \n",
+ " 0.2139 \n",
+ " 0.3801 \n",
+ " 0.4176 \n",
+ " 1 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " id_000779bfc \n",
+ " com_droga \n",
+ " 72 \n",
+ " 1 \n",
+ " df89a8e5a \n",
+ " 0.0743 \n",
+ " 0.4087 \n",
+ " 0.2991 \n",
+ " 0.0604 \n",
+ " 1.0190 \n",
+ " 0.5207 \n",
+ " 0.2341 \n",
+ " 0.3372 \n",
+ " -0.4047 \n",
+ " 0.8507 \n",
+ " -1.1520 \n",
+ " -0.4201 \n",
+ " -0.0958 \n",
+ " 0.4590 \n",
+ " 0.0803 \n",
+ " 0.2250 \n",
+ " 0.5293 \n",
+ " 0.2839 \n",
+ " -0.3494 \n",
+ " 0.2883 \n",
+ " 0.9449 \n",
+ " -0.1646 \n",
+ " -0.2657 \n",
+ " -0.3372 \n",
+ " 0.3135 \n",
+ " -0.4316 \n",
+ " 0.4773 \n",
+ " 0.2075 \n",
+ " -0.4216 \n",
+ " -0.1161 \n",
+ " -0.0499 \n",
+ " -0.2627 \n",
+ " 0.9959 \n",
+ " -0.2483 \n",
+ " 0.2655 \n",
+ " ... \n",
+ " 0.2912 \n",
+ " 0.4151 \n",
+ " -0.2840 \n",
+ " -0.3104 \n",
+ " -0.6373 \n",
+ " 0.2887 \n",
+ " -0.0765 \n",
+ " 0.2539 \n",
+ " 0.4443 \n",
+ " 0.5932 \n",
+ " 0.2031 \n",
+ " 0.7639 \n",
+ " 0.5499 \n",
+ " -0.3322 \n",
+ " -0.0977 \n",
+ " 0.4329 \n",
+ " -0.2782 \n",
+ " 0.7827 \n",
+ " 0.5934 \n",
+ " 0.3402 \n",
+ " 0.1499 \n",
+ " 0.4420 \n",
+ " 0.9366 \n",
+ " 0.8193 \n",
+ " -0.4236 \n",
+ " 0.3192 \n",
+ " -0.4265 \n",
+ " 0.7543 \n",
+ " 0.4708 \n",
+ " 0.0230 \n",
+ " 0.2957 \n",
+ " 0.4899 \n",
+ " 0.1522 \n",
+ " 0.1241 \n",
+ " 0.6077 \n",
+ " 0.7371 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " id_000a6266a \n",
+ " com_droga \n",
+ " 48 \n",
+ " 1 \n",
+ " 18bb41b2c \n",
+ " 0.6280 \n",
+ " 0.5817 \n",
+ " 1.5540 \n",
+ " -0.0764 \n",
+ " -0.0323 \n",
+ " 1.2390 \n",
+ " 0.1715 \n",
+ " 0.2155 \n",
+ " 0.0065 \n",
+ " 1.2300 \n",
+ " -0.4797 \n",
+ " -0.5631 \n",
+ " -0.0366 \n",
+ " -1.8300 \n",
+ " 0.6057 \n",
+ " -0.3278 \n",
+ " 0.6042 \n",
+ " -0.3075 \n",
+ " -0.1147 \n",
+ " -0.0570 \n",
+ " -0.0799 \n",
+ " -0.8181 \n",
+ " -1.5320 \n",
+ " 0.2307 \n",
+ " 0.4901 \n",
+ " 0.4780 \n",
+ " -1.3970 \n",
+ " 4.6240 \n",
+ " -0.0437 \n",
+ " 1.2870 \n",
+ " -1.8530 \n",
+ " 0.6069 \n",
+ " 0.4290 \n",
+ " 0.1783 \n",
+ " 0.0018 \n",
+ " ... \n",
+ " -0.0014 \n",
+ " -2.3640 \n",
+ " -0.4682 \n",
+ " 0.1210 \n",
+ " -0.5177 \n",
+ " -0.0604 \n",
+ " 0.1682 \n",
+ " -0.4436 \n",
+ " 0.4963 \n",
+ " 0.1363 \n",
+ " 0.3335 \n",
+ " 0.9760 \n",
+ " -0.0427 \n",
+ " -0.1235 \n",
+ " 0.0959 \n",
+ " 0.0690 \n",
+ " -0.9416 \n",
+ " -0.7548 \n",
+ " -0.1109 \n",
+ " -0.6272 \n",
+ " 0.3019 \n",
+ " 0.1172 \n",
+ " 0.1093 \n",
+ " -0.3113 \n",
+ " 0.3019 \n",
+ " -0.0873 \n",
+ " -0.7250 \n",
+ " -0.6297 \n",
+ " 0.6103 \n",
+ " 0.0223 \n",
+ " -1.3240 \n",
+ " -0.3174 \n",
+ " -0.6417 \n",
+ " -0.2187 \n",
+ " -1.4080 \n",
+ " 0.6931 \n",
+ " 0 \n",
+ " 1 \n",
+ " 0 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " id_0015fd391 \n",
+ " com_droga \n",
+ " 48 \n",
+ " 1 \n",
+ " 8c7f86626 \n",
+ " -0.5138 \n",
+ " -0.2491 \n",
+ " -0.2656 \n",
+ " 0.5288 \n",
+ " 4.0620 \n",
+ " -0.8095 \n",
+ " -1.9590 \n",
+ " 0.1792 \n",
+ " -0.1321 \n",
+ " -1.0600 \n",
+ " -0.8269 \n",
+ " -0.3584 \n",
+ " -0.8511 \n",
+ " -0.5844 \n",
+ " -2.5690 \n",
+ " 0.8183 \n",
+ " -0.0532 \n",
+ " -0.8554 \n",
+ " 0.1160 \n",
+ " -2.3520 \n",
+ " 2.1200 \n",
+ " -1.1580 \n",
+ " -0.7191 \n",
+ " -0.8004 \n",
+ " -1.4670 \n",
+ " -0.0107 \n",
+ " -0.8995 \n",
+ " 0.2406 \n",
+ " -0.2479 \n",
+ " -1.0890 \n",
+ " -0.7575 \n",
+ " 0.0881 \n",
+ " -2.7370 \n",
+ " 0.8745 \n",
+ " 0.5787 \n",
+ " ... \n",
+ " -2.9740 \n",
+ " -1.4930 \n",
+ " -1.6600 \n",
+ " -3.1660 \n",
+ " 0.2816 \n",
+ " -0.2990 \n",
+ " -1.1870 \n",
+ " -0.5044 \n",
+ " -1.7750 \n",
+ " -1.6120 \n",
+ " -0.9215 \n",
+ " -1.0810 \n",
+ " -3.0520 \n",
+ " -3.4470 \n",
+ " -2.7740 \n",
+ " -1.8460 \n",
+ " -0.5568 \n",
+ " -3.3960 \n",
+ " -2.9510 \n",
+ " -1.1550 \n",
+ " -3.2620 \n",
+ " -1.5390 \n",
+ " -2.4600 \n",
+ " -0.9417 \n",
+ " -1.5550 \n",
+ " 0.2431 \n",
+ " -2.0990 \n",
+ " -0.6441 \n",
+ " -5.6300 \n",
+ " -1.3780 \n",
+ " -0.8632 \n",
+ " -1.2880 \n",
+ " -1.6210 \n",
+ " -0.8784 \n",
+ " -0.3876 \n",
+ " -0.8154 \n",
+ " 0 \n",
+ " 1 \n",
+ " 0 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " id_001626bd3 \n",
+ " com_droga \n",
+ " 72 \n",
+ " 0 \n",
+ " 7cbed3131 \n",
+ " -0.3254 \n",
+ " -0.4009 \n",
+ " 0.9700 \n",
+ " 0.6919 \n",
+ " 1.4180 \n",
+ " -0.8244 \n",
+ " -0.2800 \n",
+ " -0.1498 \n",
+ " -0.8789 \n",
+ " 0.8630 \n",
+ " -0.2219 \n",
+ " -0.5121 \n",
+ " -0.9577 \n",
+ " 1.1750 \n",
+ " 0.2042 \n",
+ " 0.1970 \n",
+ " 0.1244 \n",
+ " -1.7090 \n",
+ " -0.3543 \n",
+ " -0.5160 \n",
+ " -0.3330 \n",
+ " -0.2685 \n",
+ " 0.7649 \n",
+ " 0.2057 \n",
+ " 1.3720 \n",
+ " 0.6835 \n",
+ " 0.8056 \n",
+ " -0.3754 \n",
+ " -1.2090 \n",
+ " 0.2965 \n",
+ " -0.0712 \n",
+ " 0.6389 \n",
+ " 0.6674 \n",
+ " -0.0783 \n",
+ " 1.1740 \n",
+ " ... \n",
+ " -0.5943 \n",
+ " 0.3973 \n",
+ " 0.1500 \n",
+ " 0.5178 \n",
+ " 0.5159 \n",
+ " 0.6091 \n",
+ " 0.1813 \n",
+ " -0.4249 \n",
+ " 0.7832 \n",
+ " 0.6529 \n",
+ " 0.5648 \n",
+ " 0.4817 \n",
+ " 0.0587 \n",
+ " 0.5303 \n",
+ " 0.6376 \n",
+ " -0.3966 \n",
+ " -1.4950 \n",
+ " -0.9625 \n",
+ " -0.0541 \n",
+ " 0.6273 \n",
+ " 0.4563 \n",
+ " 0.0698 \n",
+ " 0.8134 \n",
+ " 0.1924 \n",
+ " 0.6054 \n",
+ " -0.1824 \n",
+ " 0.0042 \n",
+ " 0.0048 \n",
+ " 0.6670 \n",
+ " 1.0690 \n",
+ " 0.5523 \n",
+ " -0.3031 \n",
+ " 0.1094 \n",
+ " 0.2885 \n",
+ " -0.3786 \n",
+ " 0.7125 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
5 rows × 881 columns
\n",
+ "
\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " id \n",
+ " tratamento \n",
+ " tempo \n",
+ " dose \n",
+ " droga \n",
+ " g0 \n",
+ " g1 \n",
+ " g2 \n",
+ " g3 \n",
+ " g4 \n",
+ " g5 \n",
+ " g6 \n",
+ " g7 \n",
+ " g8 \n",
+ " g9 \n",
+ " g10 \n",
+ " g11 \n",
+ " g12 \n",
+ " g13 \n",
+ " g14 \n",
+ " g15 \n",
+ " g16 \n",
+ " g17 \n",
+ " g18 \n",
+ " g19 \n",
+ " g20 \n",
+ " g21 \n",
+ " g22 \n",
+ " g23 \n",
+ " g24 \n",
+ " g25 \n",
+ " g26 \n",
+ " g27 \n",
+ " g28 \n",
+ " g29 \n",
+ " g30 \n",
+ " g31 \n",
+ " g32 \n",
+ " g33 \n",
+ " g34 \n",
+ " ... \n",
+ " c64 \n",
+ " c65 \n",
+ " c66 \n",
+ " c67 \n",
+ " c68 \n",
+ " c69 \n",
+ " c70 \n",
+ " c71 \n",
+ " c72 \n",
+ " c73 \n",
+ " c74 \n",
+ " c75 \n",
+ " c76 \n",
+ " c77 \n",
+ " c78 \n",
+ " c79 \n",
+ " c80 \n",
+ " c81 \n",
+ " c82 \n",
+ " c83 \n",
+ " c84 \n",
+ " c85 \n",
+ " c86 \n",
+ " c87 \n",
+ " c88 \n",
+ " c89 \n",
+ " c90 \n",
+ " c91 \n",
+ " c92 \n",
+ " c93 \n",
+ " c94 \n",
+ " c95 \n",
+ " c96 \n",
+ " c97 \n",
+ " c98 \n",
+ " c99 \n",
+ " tempo_24 \n",
+ " tempo_48 \n",
+ " tempo_72 \n",
+ " com_droga \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " id_000644bb2 \n",
+ " com_droga \n",
+ " 24 \n",
+ " 1 \n",
+ " b68db1d53 \n",
+ " 1.0620 \n",
+ " 0.5577 \n",
+ " -0.2479 \n",
+ " -0.6208 \n",
+ " -0.1944 \n",
+ " -1.0120 \n",
+ " -1.0220 \n",
+ " -0.0326 \n",
+ " 0.5548 \n",
+ " -0.0921 \n",
+ " 1.1830 \n",
+ " 0.1530 \n",
+ " 0.5574 \n",
+ " -0.4015 \n",
+ " 0.1789 \n",
+ " -0.6528 \n",
+ " -0.7969 \n",
+ " 0.6342 \n",
+ " 0.1778 \n",
+ " -0.3694 \n",
+ " -0.5688 \n",
+ " -1.1360 \n",
+ " -1.1880 \n",
+ " 0.6940 \n",
+ " 0.4393 \n",
+ " 0.2664 \n",
+ " 0.1907 \n",
+ " 0.1628 \n",
+ " -0.2853 \n",
+ " 0.5819 \n",
+ " 0.2934 \n",
+ " -0.5584 \n",
+ " -0.0916 \n",
+ " -0.3010 \n",
+ " -0.1537 \n",
+ " ... \n",
+ " 0.1042 \n",
+ " 0.1403 \n",
+ " 0.1758 \n",
+ " 1.2570 \n",
+ " -0.5979 \n",
+ " 1.2250 \n",
+ " -0.0553 \n",
+ " 0.7351 \n",
+ " 0.5810 \n",
+ " 0.9590 \n",
+ " 0.2427 \n",
+ " 0.0495 \n",
+ " 0.4141 \n",
+ " 0.8432 \n",
+ " 0.6162 \n",
+ " -0.7318 \n",
+ " 1.2120 \n",
+ " 0.6362 \n",
+ " -0.4427 \n",
+ " 0.1288 \n",
+ " 1.4840 \n",
+ " 0.1799 \n",
+ " 0.5367 \n",
+ " -0.1111 \n",
+ " -1.0120 \n",
+ " 0.6685 \n",
+ " 0.2862 \n",
+ " 0.2584 \n",
+ " 0.8076 \n",
+ " 0.5523 \n",
+ " -0.1912 \n",
+ " 0.6584 \n",
+ " -0.3981 \n",
+ " 0.2139 \n",
+ " 0.3801 \n",
+ " 0.4176 \n",
+ " 1 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " id_000779bfc \n",
+ " com_droga \n",
+ " 72 \n",
+ " 1 \n",
+ " df89a8e5a \n",
+ " 0.0743 \n",
+ " 0.4087 \n",
+ " 0.2991 \n",
+ " 0.0604 \n",
+ " 1.0190 \n",
+ " 0.5207 \n",
+ " 0.2341 \n",
+ " 0.3372 \n",
+ " -0.4047 \n",
+ " 0.8507 \n",
+ " -1.1520 \n",
+ " -0.4201 \n",
+ " -0.0958 \n",
+ " 0.4590 \n",
+ " 0.0803 \n",
+ " 0.2250 \n",
+ " 0.5293 \n",
+ " 0.2839 \n",
+ " -0.3494 \n",
+ " 0.2883 \n",
+ " 0.9449 \n",
+ " -0.1646 \n",
+ " -0.2657 \n",
+ " -0.3372 \n",
+ " 0.3135 \n",
+ " -0.4316 \n",
+ " 0.4773 \n",
+ " 0.2075 \n",
+ " -0.4216 \n",
+ " -0.1161 \n",
+ " -0.0499 \n",
+ " -0.2627 \n",
+ " 0.9959 \n",
+ " -0.2483 \n",
+ " 0.2655 \n",
+ " ... \n",
+ " 0.2912 \n",
+ " 0.4151 \n",
+ " -0.2840 \n",
+ " -0.3104 \n",
+ " -0.6373 \n",
+ " 0.2887 \n",
+ " -0.0765 \n",
+ " 0.2539 \n",
+ " 0.4443 \n",
+ " 0.5932 \n",
+ " 0.2031 \n",
+ " 0.7639 \n",
+ " 0.5499 \n",
+ " -0.3322 \n",
+ " -0.0977 \n",
+ " 0.4329 \n",
+ " -0.2782 \n",
+ " 0.7827 \n",
+ " 0.5934 \n",
+ " 0.3402 \n",
+ " 0.1499 \n",
+ " 0.4420 \n",
+ " 0.9366 \n",
+ " 0.8193 \n",
+ " -0.4236 \n",
+ " 0.3192 \n",
+ " -0.4265 \n",
+ " 0.7543 \n",
+ " 0.4708 \n",
+ " 0.0230 \n",
+ " 0.2957 \n",
+ " 0.4899 \n",
+ " 0.1522 \n",
+ " 0.1241 \n",
+ " 0.6077 \n",
+ " 0.7371 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " id_000a6266a \n",
+ " com_droga \n",
+ " 48 \n",
+ " 1 \n",
+ " 18bb41b2c \n",
+ " 0.6280 \n",
+ " 0.5817 \n",
+ " 1.5540 \n",
+ " -0.0764 \n",
+ " -0.0323 \n",
+ " 1.2390 \n",
+ " 0.1715 \n",
+ " 0.2155 \n",
+ " 0.0065 \n",
+ " 1.2300 \n",
+ " -0.4797 \n",
+ " -0.5631 \n",
+ " -0.0366 \n",
+ " -1.8300 \n",
+ " 0.6057 \n",
+ " -0.3278 \n",
+ " 0.6042 \n",
+ " -0.3075 \n",
+ " -0.1147 \n",
+ " -0.0570 \n",
+ " -0.0799 \n",
+ " -0.8181 \n",
+ " -1.5320 \n",
+ " 0.2307 \n",
+ " 0.4901 \n",
+ " 0.4780 \n",
+ " -1.3970 \n",
+ " 4.6240 \n",
+ " -0.0437 \n",
+ " 1.2870 \n",
+ " -1.8530 \n",
+ " 0.6069 \n",
+ " 0.4290 \n",
+ " 0.1783 \n",
+ " 0.0018 \n",
+ " ... \n",
+ " -0.0014 \n",
+ " -2.3640 \n",
+ " -0.4682 \n",
+ " 0.1210 \n",
+ " -0.5177 \n",
+ " -0.0604 \n",
+ " 0.1682 \n",
+ " -0.4436 \n",
+ " 0.4963 \n",
+ " 0.1363 \n",
+ " 0.3335 \n",
+ " 0.9760 \n",
+ " -0.0427 \n",
+ " -0.1235 \n",
+ " 0.0959 \n",
+ " 0.0690 \n",
+ " -0.9416 \n",
+ " -0.7548 \n",
+ " -0.1109 \n",
+ " -0.6272 \n",
+ " 0.3019 \n",
+ " 0.1172 \n",
+ " 0.1093 \n",
+ " -0.3113 \n",
+ " 0.3019 \n",
+ " -0.0873 \n",
+ " -0.7250 \n",
+ " -0.6297 \n",
+ " 0.6103 \n",
+ " 0.0223 \n",
+ " -1.3240 \n",
+ " -0.3174 \n",
+ " -0.6417 \n",
+ " -0.2187 \n",
+ " -1.4080 \n",
+ " 0.6931 \n",
+ " 0 \n",
+ " 1 \n",
+ " 0 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " id_0015fd391 \n",
+ " com_droga \n",
+ " 48 \n",
+ " 1 \n",
+ " 8c7f86626 \n",
+ " -0.5138 \n",
+ " -0.2491 \n",
+ " -0.2656 \n",
+ " 0.5288 \n",
+ " 4.0620 \n",
+ " -0.8095 \n",
+ " -1.9590 \n",
+ " 0.1792 \n",
+ " -0.1321 \n",
+ " -1.0600 \n",
+ " -0.8269 \n",
+ " -0.3584 \n",
+ " -0.8511 \n",
+ " -0.5844 \n",
+ " -2.5690 \n",
+ " 0.8183 \n",
+ " -0.0532 \n",
+ " -0.8554 \n",
+ " 0.1160 \n",
+ " -2.3520 \n",
+ " 2.1200 \n",
+ " -1.1580 \n",
+ " -0.7191 \n",
+ " -0.8004 \n",
+ " -1.4670 \n",
+ " -0.0107 \n",
+ " -0.8995 \n",
+ " 0.2406 \n",
+ " -0.2479 \n",
+ " -1.0890 \n",
+ " -0.7575 \n",
+ " 0.0881 \n",
+ " -2.7370 \n",
+ " 0.8745 \n",
+ " 0.5787 \n",
+ " ... \n",
+ " -2.9740 \n",
+ " -1.4930 \n",
+ " -1.6600 \n",
+ " -3.1660 \n",
+ " 0.2816 \n",
+ " -0.2990 \n",
+ " -1.1870 \n",
+ " -0.5044 \n",
+ " -1.7750 \n",
+ " -1.6120 \n",
+ " -0.9215 \n",
+ " -1.0810 \n",
+ " -3.0520 \n",
+ " -3.4470 \n",
+ " -2.7740 \n",
+ " -1.8460 \n",
+ " -0.5568 \n",
+ " -3.3960 \n",
+ " -2.9510 \n",
+ " -1.1550 \n",
+ " -3.2620 \n",
+ " -1.5390 \n",
+ " -2.4600 \n",
+ " -0.9417 \n",
+ " -1.5550 \n",
+ " 0.2431 \n",
+ " -2.0990 \n",
+ " -0.6441 \n",
+ " -5.6300 \n",
+ " -1.3780 \n",
+ " -0.8632 \n",
+ " -1.2880 \n",
+ " -1.6210 \n",
+ " -0.8784 \n",
+ " -0.3876 \n",
+ " -0.8154 \n",
+ " 0 \n",
+ " 1 \n",
+ " 0 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " id_001626bd3 \n",
+ " com_droga \n",
+ " 72 \n",
+ " 0 \n",
+ " 7cbed3131 \n",
+ " -0.3254 \n",
+ " -0.4009 \n",
+ " 0.9700 \n",
+ " 0.6919 \n",
+ " 1.4180 \n",
+ " -0.8244 \n",
+ " -0.2800 \n",
+ " -0.1498 \n",
+ " -0.8789 \n",
+ " 0.8630 \n",
+ " -0.2219 \n",
+ " -0.5121 \n",
+ " -0.9577 \n",
+ " 1.1750 \n",
+ " 0.2042 \n",
+ " 0.1970 \n",
+ " 0.1244 \n",
+ " -1.7090 \n",
+ " -0.3543 \n",
+ " -0.5160 \n",
+ " -0.3330 \n",
+ " -0.2685 \n",
+ " 0.7649 \n",
+ " 0.2057 \n",
+ " 1.3720 \n",
+ " 0.6835 \n",
+ " 0.8056 \n",
+ " -0.3754 \n",
+ " -1.2090 \n",
+ " 0.2965 \n",
+ " -0.0712 \n",
+ " 0.6389 \n",
+ " 0.6674 \n",
+ " -0.0783 \n",
+ " 1.1740 \n",
+ " ... \n",
+ " -0.5943 \n",
+ " 0.3973 \n",
+ " 0.1500 \n",
+ " 0.5178 \n",
+ " 0.5159 \n",
+ " 0.6091 \n",
+ " 0.1813 \n",
+ " -0.4249 \n",
+ " 0.7832 \n",
+ " 0.6529 \n",
+ " 0.5648 \n",
+ " 0.4817 \n",
+ " 0.0587 \n",
+ " 0.5303 \n",
+ " 0.6376 \n",
+ " -0.3966 \n",
+ " -1.4950 \n",
+ " -0.9625 \n",
+ " -0.0541 \n",
+ " 0.6273 \n",
+ " 0.4563 \n",
+ " 0.0698 \n",
+ " 0.8134 \n",
+ " 0.1924 \n",
+ " 0.6054 \n",
+ " -0.1824 \n",
+ " 0.0042 \n",
+ " 0.0048 \n",
+ " 0.6670 \n",
+ " 1.0690 \n",
+ " 0.5523 \n",
+ " -0.3031 \n",
+ " 0.1094 \n",
+ " 0.2885 \n",
+ " -0.3786 \n",
+ " 0.7125 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
5 rows × 881 columns
\n",
+ "
\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " id \n",
+ " 5-alpha_reductase_inhibitor \n",
+ " 11-beta-hsd1_inhibitor \n",
+ " acat_inhibitor \n",
+ " acetylcholine_receptor_agonist \n",
+ " acetylcholine_receptor_antagonist \n",
+ " acetylcholinesterase_inhibitor \n",
+ " adenosine_receptor_agonist \n",
+ " adenosine_receptor_antagonist \n",
+ " adenylyl_cyclase_activator \n",
+ " adrenergic_receptor_agonist \n",
+ " adrenergic_receptor_antagonist \n",
+ " akt_inhibitor \n",
+ " aldehyde_dehydrogenase_inhibitor \n",
+ " alk_inhibitor \n",
+ " ampk_activator \n",
+ " analgesic \n",
+ " androgen_receptor_agonist \n",
+ " androgen_receptor_antagonist \n",
+ " anesthetic_-_local \n",
+ " angiogenesis_inhibitor \n",
+ " angiotensin_receptor_antagonist \n",
+ " anti-inflammatory \n",
+ " antiarrhythmic \n",
+ " antibiotic \n",
+ " anticonvulsant \n",
+ " antifungal \n",
+ " antihistamine \n",
+ " antimalarial \n",
+ " antioxidant \n",
+ " antiprotozoal \n",
+ " antiviral \n",
+ " apoptosis_stimulant \n",
+ " aromatase_inhibitor \n",
+ " atm_kinase_inhibitor \n",
+ " atp-sensitive_potassium_channel_antagonist \n",
+ " atp_synthase_inhibitor \n",
+ " atpase_inhibitor \n",
+ " atr_kinase_inhibitor \n",
+ " aurora_kinase_inhibitor \n",
+ " ... \n",
+ " protein_tyrosine_kinase_inhibitor \n",
+ " radiopaque_medium \n",
+ " raf_inhibitor \n",
+ " ras_gtpase_inhibitor \n",
+ " retinoid_receptor_agonist \n",
+ " retinoid_receptor_antagonist \n",
+ " rho_associated_kinase_inhibitor \n",
+ " ribonucleoside_reductase_inhibitor \n",
+ " rna_polymerase_inhibitor \n",
+ " serotonin_receptor_agonist \n",
+ " serotonin_receptor_antagonist \n",
+ " serotonin_reuptake_inhibitor \n",
+ " sigma_receptor_agonist \n",
+ " sigma_receptor_antagonist \n",
+ " smoothened_receptor_antagonist \n",
+ " sodium_channel_inhibitor \n",
+ " sphingosine_receptor_agonist \n",
+ " src_inhibitor \n",
+ " steroid \n",
+ " syk_inhibitor \n",
+ " tachykinin_antagonist \n",
+ " tgf-beta_receptor_inhibitor \n",
+ " thrombin_inhibitor \n",
+ " thymidylate_synthase_inhibitor \n",
+ " tlr_agonist \n",
+ " tlr_antagonist \n",
+ " tnf_inhibitor \n",
+ " topoisomerase_inhibitor \n",
+ " transient_receptor_potential_channel_antagonist \n",
+ " tropomyosin_receptor_kinase_inhibitor \n",
+ " trpv_agonist \n",
+ " trpv_antagonist \n",
+ " tubulin_inhibitor \n",
+ " tyrosine_kinase_inhibitor \n",
+ " ubiquitin_specific_protease_inhibitor \n",
+ " vegfr_inhibitor \n",
+ " vitamin_b \n",
+ " vitamin_d_receptor_agonist \n",
+ " wnt_inhibitor \n",
+ " qtd_moa \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " id_000644bb2 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " id_000779bfc \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " id_000a6266a \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 3 \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " id_0015fd391 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " id_001626bd3 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
5 rows × 208 columns
\n",
+ "
\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " id \n",
+ " 5-alpha_reductase_inhibitor \n",
+ " 11-beta-hsd1_inhibitor \n",
+ " acat_inhibitor \n",
+ " acetylcholine_receptor_agonist \n",
+ " acetylcholine_receptor_antagonist \n",
+ " acetylcholinesterase_inhibitor \n",
+ " adenosine_receptor_agonist \n",
+ " adenosine_receptor_antagonist \n",
+ " adenylyl_cyclase_activator \n",
+ " adrenergic_receptor_agonist \n",
+ " adrenergic_receptor_antagonist \n",
+ " akt_inhibitor \n",
+ " aldehyde_dehydrogenase_inhibitor \n",
+ " alk_inhibitor \n",
+ " ampk_activator \n",
+ " analgesic \n",
+ " androgen_receptor_agonist \n",
+ " androgen_receptor_antagonist \n",
+ " anesthetic_-_local \n",
+ " angiogenesis_inhibitor \n",
+ " angiotensin_receptor_antagonist \n",
+ " anti-inflammatory \n",
+ " antiarrhythmic \n",
+ " antibiotic \n",
+ " anticonvulsant \n",
+ " antifungal \n",
+ " antihistamine \n",
+ " antimalarial \n",
+ " antioxidant \n",
+ " antiprotozoal \n",
+ " antiviral \n",
+ " apoptosis_stimulant \n",
+ " aromatase_inhibitor \n",
+ " atm_kinase_inhibitor \n",
+ " atp-sensitive_potassium_channel_antagonist \n",
+ " atp_synthase_inhibitor \n",
+ " atpase_inhibitor \n",
+ " atr_kinase_inhibitor \n",
+ " aurora_kinase_inhibitor \n",
+ " ... \n",
+ " radiopaque_medium \n",
+ " raf_inhibitor \n",
+ " ras_gtpase_inhibitor \n",
+ " retinoid_receptor_agonist \n",
+ " retinoid_receptor_antagonist \n",
+ " rho_associated_kinase_inhibitor \n",
+ " ribonucleoside_reductase_inhibitor \n",
+ " rna_polymerase_inhibitor \n",
+ " serotonin_receptor_agonist \n",
+ " serotonin_receptor_antagonist \n",
+ " serotonin_reuptake_inhibitor \n",
+ " sigma_receptor_agonist \n",
+ " sigma_receptor_antagonist \n",
+ " smoothened_receptor_antagonist \n",
+ " sodium_channel_inhibitor \n",
+ " sphingosine_receptor_agonist \n",
+ " src_inhibitor \n",
+ " steroid \n",
+ " syk_inhibitor \n",
+ " tachykinin_antagonist \n",
+ " tgf-beta_receptor_inhibitor \n",
+ " thrombin_inhibitor \n",
+ " thymidylate_synthase_inhibitor \n",
+ " tlr_agonist \n",
+ " tlr_antagonist \n",
+ " tnf_inhibitor \n",
+ " topoisomerase_inhibitor \n",
+ " transient_receptor_potential_channel_antagonist \n",
+ " tropomyosin_receptor_kinase_inhibitor \n",
+ " trpv_agonist \n",
+ " trpv_antagonist \n",
+ " tubulin_inhibitor \n",
+ " tyrosine_kinase_inhibitor \n",
+ " ubiquitin_specific_protease_inhibitor \n",
+ " vegfr_inhibitor \n",
+ " vitamin_b \n",
+ " vitamin_d_receptor_agonist \n",
+ " wnt_inhibitor \n",
+ " qtd_moa \n",
+ " tem_moa \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " id_000644bb2 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " True \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " id_000779bfc \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " False \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " id_000a6266a \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 3 \n",
+ " True \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " id_0015fd391 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " False \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " id_001626bd3 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " True \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
5 rows × 209 columns
\n",
+ "
\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " id \n",
+ " 5-alpha_reductase_inhibitor \n",
+ " 11-beta-hsd1_inhibitor \n",
+ " acat_inhibitor \n",
+ " acetylcholine_receptor_agonist \n",
+ " acetylcholine_receptor_antagonist \n",
+ " acetylcholinesterase_inhibitor \n",
+ " adenosine_receptor_agonist \n",
+ " adenosine_receptor_antagonist \n",
+ " adenylyl_cyclase_activator \n",
+ " adrenergic_receptor_agonist \n",
+ " adrenergic_receptor_antagonist \n",
+ " akt_inhibitor \n",
+ " aldehyde_dehydrogenase_inhibitor \n",
+ " alk_inhibitor \n",
+ " ampk_activator \n",
+ " analgesic \n",
+ " androgen_receptor_agonist \n",
+ " androgen_receptor_antagonist \n",
+ " anesthetic_-_local \n",
+ " angiogenesis_inhibitor \n",
+ " angiotensin_receptor_antagonist \n",
+ " anti-inflammatory \n",
+ " antiarrhythmic \n",
+ " antibiotic \n",
+ " anticonvulsant \n",
+ " antifungal \n",
+ " antihistamine \n",
+ " antimalarial \n",
+ " antioxidant \n",
+ " antiprotozoal \n",
+ " antiviral \n",
+ " apoptosis_stimulant \n",
+ " aromatase_inhibitor \n",
+ " atm_kinase_inhibitor \n",
+ " atp-sensitive_potassium_channel_antagonist \n",
+ " atp_synthase_inhibitor \n",
+ " atpase_inhibitor \n",
+ " atr_kinase_inhibitor \n",
+ " aurora_kinase_inhibitor \n",
+ " ... \n",
+ " raf_inhibitor \n",
+ " ras_gtpase_inhibitor \n",
+ " retinoid_receptor_agonist \n",
+ " retinoid_receptor_antagonist \n",
+ " rho_associated_kinase_inhibitor \n",
+ " ribonucleoside_reductase_inhibitor \n",
+ " rna_polymerase_inhibitor \n",
+ " serotonin_receptor_agonist \n",
+ " serotonin_receptor_antagonist \n",
+ " serotonin_reuptake_inhibitor \n",
+ " sigma_receptor_agonist \n",
+ " sigma_receptor_antagonist \n",
+ " smoothened_receptor_antagonist \n",
+ " sodium_channel_inhibitor \n",
+ " sphingosine_receptor_agonist \n",
+ " src_inhibitor \n",
+ " steroid \n",
+ " syk_inhibitor \n",
+ " tachykinin_antagonist \n",
+ " tgf-beta_receptor_inhibitor \n",
+ " thrombin_inhibitor \n",
+ " thymidylate_synthase_inhibitor \n",
+ " tlr_agonist \n",
+ " tlr_antagonist \n",
+ " tnf_inhibitor \n",
+ " topoisomerase_inhibitor \n",
+ " transient_receptor_potential_channel_antagonist \n",
+ " tropomyosin_receptor_kinase_inhibitor \n",
+ " trpv_agonist \n",
+ " trpv_antagonist \n",
+ " tubulin_inhibitor \n",
+ " tyrosine_kinase_inhibitor \n",
+ " ubiquitin_specific_protease_inhibitor \n",
+ " vegfr_inhibitor \n",
+ " vitamin_b \n",
+ " vitamin_d_receptor_agonist \n",
+ " wnt_inhibitor \n",
+ " qtd_moa \n",
+ " tem_moa \n",
+ " atv_moa \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " id_000644bb2 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " True \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " id_000779bfc \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " False \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " id_000a6266a \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 3 \n",
+ " True \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " id_0015fd391 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " False \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " id_001626bd3 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " True \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
5 rows × 210 columns
\n",
+ "
\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " id \n",
+ " 5-alpha_reductase_inhibitor \n",
+ " 11-beta-hsd1_inhibitor \n",
+ " acat_inhibitor \n",
+ " acetylcholine_receptor_agonist \n",
+ " acetylcholine_receptor_antagonist \n",
+ " acetylcholinesterase_inhibitor \n",
+ " adenosine_receptor_agonist \n",
+ " adenosine_receptor_antagonist \n",
+ " adenylyl_cyclase_activator \n",
+ " adrenergic_receptor_agonist \n",
+ " adrenergic_receptor_antagonist \n",
+ " akt_inhibitor \n",
+ " aldehyde_dehydrogenase_inhibitor \n",
+ " alk_inhibitor \n",
+ " ampk_activator \n",
+ " analgesic \n",
+ " androgen_receptor_agonist \n",
+ " androgen_receptor_antagonist \n",
+ " anesthetic_-_local \n",
+ " angiogenesis_inhibitor \n",
+ " angiotensin_receptor_antagonist \n",
+ " anti-inflammatory \n",
+ " antiarrhythmic \n",
+ " antibiotic \n",
+ " anticonvulsant \n",
+ " antifungal \n",
+ " antihistamine \n",
+ " antimalarial \n",
+ " antioxidant \n",
+ " antiprotozoal \n",
+ " antiviral \n",
+ " apoptosis_stimulant \n",
+ " aromatase_inhibitor \n",
+ " atm_kinase_inhibitor \n",
+ " atp-sensitive_potassium_channel_antagonist \n",
+ " atp_synthase_inhibitor \n",
+ " atpase_inhibitor \n",
+ " atr_kinase_inhibitor \n",
+ " aurora_kinase_inhibitor \n",
+ " ... \n",
+ " radiopaque_medium \n",
+ " raf_inhibitor \n",
+ " ras_gtpase_inhibitor \n",
+ " retinoid_receptor_agonist \n",
+ " retinoid_receptor_antagonist \n",
+ " rho_associated_kinase_inhibitor \n",
+ " ribonucleoside_reductase_inhibitor \n",
+ " rna_polymerase_inhibitor \n",
+ " serotonin_receptor_agonist \n",
+ " serotonin_receptor_antagonist \n",
+ " serotonin_reuptake_inhibitor \n",
+ " sigma_receptor_agonist \n",
+ " sigma_receptor_antagonist \n",
+ " smoothened_receptor_antagonist \n",
+ " sodium_channel_inhibitor \n",
+ " sphingosine_receptor_agonist \n",
+ " src_inhibitor \n",
+ " steroid \n",
+ " syk_inhibitor \n",
+ " tachykinin_antagonist \n",
+ " tgf-beta_receptor_inhibitor \n",
+ " thrombin_inhibitor \n",
+ " thymidylate_synthase_inhibitor \n",
+ " tlr_agonist \n",
+ " tlr_antagonist \n",
+ " tnf_inhibitor \n",
+ " topoisomerase_inhibitor \n",
+ " transient_receptor_potential_channel_antagonist \n",
+ " tropomyosin_receptor_kinase_inhibitor \n",
+ " trpv_agonist \n",
+ " trpv_antagonist \n",
+ " tubulin_inhibitor \n",
+ " tyrosine_kinase_inhibitor \n",
+ " ubiquitin_specific_protease_inhibitor \n",
+ " vegfr_inhibitor \n",
+ " vitamin_b \n",
+ " vitamin_d_receptor_agonist \n",
+ " wnt_inhibitor \n",
+ " qtd_moa \n",
+ " atv_moa \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " id_000644bb2 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " id_000779bfc \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " id_000a6266a \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 3 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " id_0015fd391 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " id_001626bd3 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " ... \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
5 rows × 209 columns
\n",
+ "
Junção das Bases "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "KMtPY2o7o4VS"
+ },
+ "source": [
+ "Já transformamos todas as colunas que queríamos, agora vamos juntar as duas bases para conseguirmos ter uma melhor análise e cruzar melhor as colunas para tirar insights"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 244
+ },
+ "id": "SxJVtYtXnBMf",
+ "outputId": "d1c352a0-5530-41fe-b104-d42eb7649094"
+ },
+ "source": [
+ "df = pd.merge(df_experimentos, df_resultados[['id','qtd_moa', 'atv_moa',]], on='id')\n",
+ "df.head()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " id \n",
+ " tratamento \n",
+ " tempo \n",
+ " dose \n",
+ " droga \n",
+ " g0 \n",
+ " g1 \n",
+ " g2 \n",
+ " g3 \n",
+ " g4 \n",
+ " g5 \n",
+ " g6 \n",
+ " g7 \n",
+ " g8 \n",
+ " g9 \n",
+ " g10 \n",
+ " g11 \n",
+ " g12 \n",
+ " g13 \n",
+ " g14 \n",
+ " g15 \n",
+ " g16 \n",
+ " g17 \n",
+ " g18 \n",
+ " g19 \n",
+ " g20 \n",
+ " g21 \n",
+ " g22 \n",
+ " g23 \n",
+ " g24 \n",
+ " g25 \n",
+ " g26 \n",
+ " g27 \n",
+ " g28 \n",
+ " g29 \n",
+ " g30 \n",
+ " g31 \n",
+ " g32 \n",
+ " g33 \n",
+ " g34 \n",
+ " ... \n",
+ " c66 \n",
+ " c67 \n",
+ " c68 \n",
+ " c69 \n",
+ " c70 \n",
+ " c71 \n",
+ " c72 \n",
+ " c73 \n",
+ " c74 \n",
+ " c75 \n",
+ " c76 \n",
+ " c77 \n",
+ " c78 \n",
+ " c79 \n",
+ " c80 \n",
+ " c81 \n",
+ " c82 \n",
+ " c83 \n",
+ " c84 \n",
+ " c85 \n",
+ " c86 \n",
+ " c87 \n",
+ " c88 \n",
+ " c89 \n",
+ " c90 \n",
+ " c91 \n",
+ " c92 \n",
+ " c93 \n",
+ " c94 \n",
+ " c95 \n",
+ " c96 \n",
+ " c97 \n",
+ " c98 \n",
+ " c99 \n",
+ " tempo_24 \n",
+ " tempo_48 \n",
+ " tempo_72 \n",
+ " com_droga \n",
+ " qtd_moa \n",
+ " atv_moa \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 0 \n",
+ " id_000644bb2 \n",
+ " com_droga \n",
+ " 24 \n",
+ " 1 \n",
+ " b68db1d53 \n",
+ " 1.0620 \n",
+ " 0.5577 \n",
+ " -0.2479 \n",
+ " -0.6208 \n",
+ " -0.1944 \n",
+ " -1.0120 \n",
+ " -1.0220 \n",
+ " -0.0326 \n",
+ " 0.5548 \n",
+ " -0.0921 \n",
+ " 1.1830 \n",
+ " 0.1530 \n",
+ " 0.5574 \n",
+ " -0.4015 \n",
+ " 0.1789 \n",
+ " -0.6528 \n",
+ " -0.7969 \n",
+ " 0.6342 \n",
+ " 0.1778 \n",
+ " -0.3694 \n",
+ " -0.5688 \n",
+ " -1.1360 \n",
+ " -1.1880 \n",
+ " 0.6940 \n",
+ " 0.4393 \n",
+ " 0.2664 \n",
+ " 0.1907 \n",
+ " 0.1628 \n",
+ " -0.2853 \n",
+ " 0.5819 \n",
+ " 0.2934 \n",
+ " -0.5584 \n",
+ " -0.0916 \n",
+ " -0.3010 \n",
+ " -0.1537 \n",
+ " ... \n",
+ " 0.1758 \n",
+ " 1.2570 \n",
+ " -0.5979 \n",
+ " 1.2250 \n",
+ " -0.0553 \n",
+ " 0.7351 \n",
+ " 0.5810 \n",
+ " 0.9590 \n",
+ " 0.2427 \n",
+ " 0.0495 \n",
+ " 0.4141 \n",
+ " 0.8432 \n",
+ " 0.6162 \n",
+ " -0.7318 \n",
+ " 1.2120 \n",
+ " 0.6362 \n",
+ " -0.4427 \n",
+ " 0.1288 \n",
+ " 1.4840 \n",
+ " 0.1799 \n",
+ " 0.5367 \n",
+ " -0.1111 \n",
+ " -1.0120 \n",
+ " 0.6685 \n",
+ " 0.2862 \n",
+ " 0.2584 \n",
+ " 0.8076 \n",
+ " 0.5523 \n",
+ " -0.1912 \n",
+ " 0.6584 \n",
+ " -0.3981 \n",
+ " 0.2139 \n",
+ " 0.3801 \n",
+ " 0.4176 \n",
+ " 1 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 1 \n",
+ " id_000779bfc \n",
+ " com_droga \n",
+ " 72 \n",
+ " 1 \n",
+ " df89a8e5a \n",
+ " 0.0743 \n",
+ " 0.4087 \n",
+ " 0.2991 \n",
+ " 0.0604 \n",
+ " 1.0190 \n",
+ " 0.5207 \n",
+ " 0.2341 \n",
+ " 0.3372 \n",
+ " -0.4047 \n",
+ " 0.8507 \n",
+ " -1.1520 \n",
+ " -0.4201 \n",
+ " -0.0958 \n",
+ " 0.4590 \n",
+ " 0.0803 \n",
+ " 0.2250 \n",
+ " 0.5293 \n",
+ " 0.2839 \n",
+ " -0.3494 \n",
+ " 0.2883 \n",
+ " 0.9449 \n",
+ " -0.1646 \n",
+ " -0.2657 \n",
+ " -0.3372 \n",
+ " 0.3135 \n",
+ " -0.4316 \n",
+ " 0.4773 \n",
+ " 0.2075 \n",
+ " -0.4216 \n",
+ " -0.1161 \n",
+ " -0.0499 \n",
+ " -0.2627 \n",
+ " 0.9959 \n",
+ " -0.2483 \n",
+ " 0.2655 \n",
+ " ... \n",
+ " -0.2840 \n",
+ " -0.3104 \n",
+ " -0.6373 \n",
+ " 0.2887 \n",
+ " -0.0765 \n",
+ " 0.2539 \n",
+ " 0.4443 \n",
+ " 0.5932 \n",
+ " 0.2031 \n",
+ " 0.7639 \n",
+ " 0.5499 \n",
+ " -0.3322 \n",
+ " -0.0977 \n",
+ " 0.4329 \n",
+ " -0.2782 \n",
+ " 0.7827 \n",
+ " 0.5934 \n",
+ " 0.3402 \n",
+ " 0.1499 \n",
+ " 0.4420 \n",
+ " 0.9366 \n",
+ " 0.8193 \n",
+ " -0.4236 \n",
+ " 0.3192 \n",
+ " -0.4265 \n",
+ " 0.7543 \n",
+ " 0.4708 \n",
+ " 0.0230 \n",
+ " 0.2957 \n",
+ " 0.4899 \n",
+ " 0.1522 \n",
+ " 0.1241 \n",
+ " 0.6077 \n",
+ " 0.7371 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " 1 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 2 \n",
+ " id_000a6266a \n",
+ " com_droga \n",
+ " 48 \n",
+ " 1 \n",
+ " 18bb41b2c \n",
+ " 0.6280 \n",
+ " 0.5817 \n",
+ " 1.5540 \n",
+ " -0.0764 \n",
+ " -0.0323 \n",
+ " 1.2390 \n",
+ " 0.1715 \n",
+ " 0.2155 \n",
+ " 0.0065 \n",
+ " 1.2300 \n",
+ " -0.4797 \n",
+ " -0.5631 \n",
+ " -0.0366 \n",
+ " -1.8300 \n",
+ " 0.6057 \n",
+ " -0.3278 \n",
+ " 0.6042 \n",
+ " -0.3075 \n",
+ " -0.1147 \n",
+ " -0.0570 \n",
+ " -0.0799 \n",
+ " -0.8181 \n",
+ " -1.5320 \n",
+ " 0.2307 \n",
+ " 0.4901 \n",
+ " 0.4780 \n",
+ " -1.3970 \n",
+ " 4.6240 \n",
+ " -0.0437 \n",
+ " 1.2870 \n",
+ " -1.8530 \n",
+ " 0.6069 \n",
+ " 0.4290 \n",
+ " 0.1783 \n",
+ " 0.0018 \n",
+ " ... \n",
+ " -0.4682 \n",
+ " 0.1210 \n",
+ " -0.5177 \n",
+ " -0.0604 \n",
+ " 0.1682 \n",
+ " -0.4436 \n",
+ " 0.4963 \n",
+ " 0.1363 \n",
+ " 0.3335 \n",
+ " 0.9760 \n",
+ " -0.0427 \n",
+ " -0.1235 \n",
+ " 0.0959 \n",
+ " 0.0690 \n",
+ " -0.9416 \n",
+ " -0.7548 \n",
+ " -0.1109 \n",
+ " -0.6272 \n",
+ " 0.3019 \n",
+ " 0.1172 \n",
+ " 0.1093 \n",
+ " -0.3113 \n",
+ " 0.3019 \n",
+ " -0.0873 \n",
+ " -0.7250 \n",
+ " -0.6297 \n",
+ " 0.6103 \n",
+ " 0.0223 \n",
+ " -1.3240 \n",
+ " -0.3174 \n",
+ " -0.6417 \n",
+ " -0.2187 \n",
+ " -1.4080 \n",
+ " 0.6931 \n",
+ " 0 \n",
+ " 1 \n",
+ " 0 \n",
+ " 1 \n",
+ " 3 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 3 \n",
+ " id_0015fd391 \n",
+ " com_droga \n",
+ " 48 \n",
+ " 1 \n",
+ " 8c7f86626 \n",
+ " -0.5138 \n",
+ " -0.2491 \n",
+ " -0.2656 \n",
+ " 0.5288 \n",
+ " 4.0620 \n",
+ " -0.8095 \n",
+ " -1.9590 \n",
+ " 0.1792 \n",
+ " -0.1321 \n",
+ " -1.0600 \n",
+ " -0.8269 \n",
+ " -0.3584 \n",
+ " -0.8511 \n",
+ " -0.5844 \n",
+ " -2.5690 \n",
+ " 0.8183 \n",
+ " -0.0532 \n",
+ " -0.8554 \n",
+ " 0.1160 \n",
+ " -2.3520 \n",
+ " 2.1200 \n",
+ " -1.1580 \n",
+ " -0.7191 \n",
+ " -0.8004 \n",
+ " -1.4670 \n",
+ " -0.0107 \n",
+ " -0.8995 \n",
+ " 0.2406 \n",
+ " -0.2479 \n",
+ " -1.0890 \n",
+ " -0.7575 \n",
+ " 0.0881 \n",
+ " -2.7370 \n",
+ " 0.8745 \n",
+ " 0.5787 \n",
+ " ... \n",
+ " -1.6600 \n",
+ " -3.1660 \n",
+ " 0.2816 \n",
+ " -0.2990 \n",
+ " -1.1870 \n",
+ " -0.5044 \n",
+ " -1.7750 \n",
+ " -1.6120 \n",
+ " -0.9215 \n",
+ " -1.0810 \n",
+ " -3.0520 \n",
+ " -3.4470 \n",
+ " -2.7740 \n",
+ " -1.8460 \n",
+ " -0.5568 \n",
+ " -3.3960 \n",
+ " -2.9510 \n",
+ " -1.1550 \n",
+ " -3.2620 \n",
+ " -1.5390 \n",
+ " -2.4600 \n",
+ " -0.9417 \n",
+ " -1.5550 \n",
+ " 0.2431 \n",
+ " -2.0990 \n",
+ " -0.6441 \n",
+ " -5.6300 \n",
+ " -1.3780 \n",
+ " -0.8632 \n",
+ " -1.2880 \n",
+ " -1.6210 \n",
+ " -0.8784 \n",
+ " -0.3876 \n",
+ " -0.8154 \n",
+ " 0 \n",
+ " 1 \n",
+ " 0 \n",
+ " 1 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 4 \n",
+ " id_001626bd3 \n",
+ " com_droga \n",
+ " 72 \n",
+ " 0 \n",
+ " 7cbed3131 \n",
+ " -0.3254 \n",
+ " -0.4009 \n",
+ " 0.9700 \n",
+ " 0.6919 \n",
+ " 1.4180 \n",
+ " -0.8244 \n",
+ " -0.2800 \n",
+ " -0.1498 \n",
+ " -0.8789 \n",
+ " 0.8630 \n",
+ " -0.2219 \n",
+ " -0.5121 \n",
+ " -0.9577 \n",
+ " 1.1750 \n",
+ " 0.2042 \n",
+ " 0.1970 \n",
+ " 0.1244 \n",
+ " -1.7090 \n",
+ " -0.3543 \n",
+ " -0.5160 \n",
+ " -0.3330 \n",
+ " -0.2685 \n",
+ " 0.7649 \n",
+ " 0.2057 \n",
+ " 1.3720 \n",
+ " 0.6835 \n",
+ " 0.8056 \n",
+ " -0.3754 \n",
+ " -1.2090 \n",
+ " 0.2965 \n",
+ " -0.0712 \n",
+ " 0.6389 \n",
+ " 0.6674 \n",
+ " -0.0783 \n",
+ " 1.1740 \n",
+ " ... \n",
+ " 0.1500 \n",
+ " 0.5178 \n",
+ " 0.5159 \n",
+ " 0.6091 \n",
+ " 0.1813 \n",
+ " -0.4249 \n",
+ " 0.7832 \n",
+ " 0.6529 \n",
+ " 0.5648 \n",
+ " 0.4817 \n",
+ " 0.0587 \n",
+ " 0.5303 \n",
+ " 0.6376 \n",
+ " -0.3966 \n",
+ " -1.4950 \n",
+ " -0.9625 \n",
+ " -0.0541 \n",
+ " 0.6273 \n",
+ " 0.4563 \n",
+ " 0.0698 \n",
+ " 0.8134 \n",
+ " 0.1924 \n",
+ " 0.6054 \n",
+ " -0.1824 \n",
+ " 0.0042 \n",
+ " 0.0048 \n",
+ " 0.6670 \n",
+ " 1.0690 \n",
+ " 0.5523 \n",
+ " -0.3031 \n",
+ " 0.1094 \n",
+ " 0.2885 \n",
+ " -0.3786 \n",
+ " 0.7125 \n",
+ " 0 \n",
+ " 0 \n",
+ " 1 \n",
+ " 1 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
5 rows × 883 columns
\n",
+ "
***Levantamento de Hipóteses*** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "BSV3OnDypmEc"
+ },
+ "source": [
+ "Após conhecer a nossa base e fazer as transformações necessárias, vamos levantar algumas hipóteses para tentar válidar elas ou não.\n",
+ "\n",
+ "1. Todos os experimentos que foram tratados com com_controle não houve ativação do MoA\n",
+ "2. Dos experimentos que foram tratados com droga, 20% não ativaram o MoA\n",
+ "3. Metade dos experimentos que receberam a primeira dose não ativaram o MoA\n",
+ "4. 90% dos experimentos tratados com droga que receberam a segunda dose ativaram o MoA\n",
+ "5. A droga mais utilizada nos experimentos ativa o MoA 4 vezes, pelo menos uma vez\n",
+ "6. A droga mais utilizada nos experimentos não precisa tomar a segunda dose\n",
+ "7. Os experimentos que tiveram 24 horas de espera só ativaram o MoA 50% dos casos.\n",
+ "8. Os experimentos que foram tratados com droga, 75% ativa com a espera de 72 horas.\n",
+ "\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "mrdZDxtenZb9"
+ },
+ "source": [
+ "Agora que criamos nossas hipoteses, vamos verificar elas."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "6153pUybjXPT"
+ },
+ "source": [
+ "##1. Todos os experimentos que foram tratados com com_controle não houve ativação do MoA"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "PsZYlRi9ud4W"
+ },
+ "source": [
+ "Vamos verificar todos que que não foram utilizados droga e ver se tiveram MoA. Em tese, `com_controle` não deveria ter ativado o MoA"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "VJ4l3xG5pwoe",
+ "outputId": "b6e19436-84e7-4750-cc8d-7215fbaaf6fa"
+ },
+ "source": [
+ "df.query('com_droga == 0')['atv_moa'].value_counts()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "0 1866\n",
+ "Name: atv_moa, dtype: int64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 194
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "S-8POPfsuwcc"
+ },
+ "source": [
+ "Conforme previsto, realmente nenhum com controle ativou um MoA\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "MD0CAW1djlBf"
+ },
+ "source": [
+ "##2. Dos experimentos que foram tratados com droga, 20% não ativaram o MoA"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Pyx2jn24u3h3"
+ },
+ "source": [
+ "Vamos ver se os tratados com droga não ativaram o MoA"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "XxJ5Ldm00_Fa",
+ "outputId": "29e2b047-0f24-4e1a-bdd9-75cb06000441"
+ },
+ "source": [
+ "df.query('com_droga == 1')['atv_moa'].value_counts(normalize=True)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "1 0.658238\n",
+ "0 0.341762\n",
+ "Name: atv_moa, dtype: float64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 269
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "CY1qSJzOufLE",
+ "outputId": "7b065194-3a91-41e2-8ae3-585557c502ac"
+ },
+ "source": [
+ "droga_atv = df.query('com_droga == 1')['atv_moa'].value_counts()\n",
+ "\n",
+ "plt.style.use('seaborn')\n",
+ "plt.figure(figsize=(8,6))\n",
+ "ax = droga_atv.plot.bar(color=['grey','#092A32'])\n",
+ "plt.title('Tratados com Droga mas Não Ativaram',fontdict={'fontsize':18, 'fontweight': 'bold'},)\n",
+ "plt.xlabel('Ação') \n",
+ "plt.ylabel('Quantidade')\n",
+ "plt.xticks(rotation=0)\n",
+ "plt.legend(['Moa Ativou'])\n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "jmVbTRJgjuEd"
+ },
+ "source": [
+ "34% dos que receberam o tratamento com droga não ativaram o MoA"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "jXYcS3enj-Qd"
+ },
+ "source": [
+ "##3. Metade dos experimentos que receberam a primeira dose não ativaram o MoA"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "8IBHdR7nvExW",
+ "outputId": "7b892fe4-de2c-47e7-b877-071cf0aa44eb"
+ },
+ "source": [
+ "df.query('com_droga == 1 and dose == 1')['atv_moa'].value_counts(normalize=True)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "1 0.658449\n",
+ "0 0.341551\n",
+ "Name: atv_moa, dtype: float64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 270
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 407
+ },
+ "id": "ff4WoD3i2KRE",
+ "outputId": "5749d75e-a82d-4517-c877-e916093aaefd"
+ },
+ "source": [
+ "droga_d1 = df.query('com_droga == 1 and dose == 1')['atv_moa'].value_counts()\n",
+ "\n",
+ "plt.style.use('seaborn')\n",
+ "plt.figure(figsize=(8,6))\n",
+ "droga_d1.plot.bar(color=['#092A32','grey',])\n",
+ "plt.title('Experimentos que tomaram a Primeira Dose e Ativou algum MoA',fontdict={'fontsize':18, 'fontweight': 'bold'},)\n",
+ "plt.xlabel('Ação') \n",
+ "plt.ylabel('Quantidade')\n",
+ "plt.xticks(rotation=0)\n",
+ "plt.legend(['MoA Ativou'])\n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "zBLf1vLGkMSE"
+ },
+ "source": [
+ "Neste caso, apenas 34% dos experimentos receberam a primeira dose, não ativou o MoA"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "pYcuxNBPki4v"
+ },
+ "source": [
+ "##4. 90% dos experimentos que foram tratados com droga que receberam a segunda dose ativaram o MoA"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "K2gqgl12uqSX",
+ "outputId": "e130a300-1e09-4f56-c6ed-518736278993"
+ },
+ "source": [
+ "df.query('com_droga == 1 and dose == 0')['atv_moa'].value_counts(normalize=True)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "1 0.658017\n",
+ "0 0.341983\n",
+ "Name: atv_moa, dtype: float64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 272
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 407
+ },
+ "id": "Dm1gAaMY2KvE",
+ "outputId": "3b7d3e05-a2b2-4132-a2a1-9dc67af29aac"
+ },
+ "source": [
+ "droga_d2 = df.query('com_droga == 1 and dose == 0')['atv_moa'].value_counts()\n",
+ "\n",
+ "plt.style.use('seaborn')\n",
+ "plt.figure(figsize=(8,6))\n",
+ "droga_d2.plot.bar(color=['grey','#092A32'])\n",
+ "plt.title('Experimentos que tomaram a Segunda Dose e Não Ativou o MoA',fontdict={'fontsize':18, 'fontweight': 'bold'},)\n",
+ "plt.xlabel('Ação') \n",
+ "plt.ylabel('Quantidade')\n",
+ "plt.xticks(rotation=0)\n",
+ "plt.legend(['Moa Ativou'])\n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "tFEKd87elEe0"
+ },
+ "source": [
+ "Apenas 65% dos experimentos que foram tratados com droga que receberam a segunda dose ativaram o MoA"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "JO0brVEClXDP"
+ },
+ "source": [
+ "##5. A droga mais utilizada nos experimentos ativa o MoA 4 vezes pelo menos uma vez"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "WHOjtED_ukG0",
+ "outputId": "572bdcdd-e167-4915-b160-75c25ee6cbc2"
+ },
+ "source": [
+ "df.groupby(by='droga')['qtd_moa'].sum().sort_values(ascending = False)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "droga\n",
+ "87d714366 1436\n",
+ "d50f18348 558\n",
+ "9f80f3f77 246\n",
+ "8b87a7a83 203\n",
+ "5628cb3ee 202\n",
+ " ... \n",
+ "8be20f208 0\n",
+ "8c25111b8 0\n",
+ "8c48a42be 0\n",
+ "8c7b401c2 0\n",
+ "00199ff52 0\n",
+ "Name: qtd_moa, Length: 3289, dtype: int64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 201
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "cXYIpiIs4z6a",
+ "outputId": "cbff7d21-e8a4-42e5-d4e9-8283972c5452"
+ },
+ "source": [
+ "df.query('droga == \"87d714366\"')['qtd_moa'].sort_values(ascending = False)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "23802 2\n",
+ "7647 2\n",
+ "8425 2\n",
+ "8404 2\n",
+ "8359 2\n",
+ " ..\n",
+ "16260 2\n",
+ "16258 2\n",
+ "16246 2\n",
+ "16235 2\n",
+ "16 2\n",
+ "Name: qtd_moa, Length: 718, dtype: int64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 202
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "zdpARoHh8jiG"
+ },
+ "source": [
+ "A droga mais utilizada nos experimentos ativa o 2 MoAs **todas** das vezes."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "H9u7NxWA8qGH"
+ },
+ "source": [
+ "##6. A droga mais utilizada nos experimentos não precisa tomar a segunda dose"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "nqorggtm4kX7",
+ "outputId": "38cd5d8f-d19f-4a3c-9f7a-497603c4a5c6"
+ },
+ "source": [
+ "df.query('droga == \"87d714366\"')['dose'].value_counts()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "1 375\n",
+ "0 343\n",
+ "Name: dose, dtype: int64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 203
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "ozSiRWRhluzU"
+ },
+ "source": [
+ "A droga mais utilizada tem casos que não ativou o MoA."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "GTo4L6jznFSl"
+ },
+ "source": [
+ "##7. Os experimentos que tiveram 24 horas de espera só ativaram o MoA 50% dos casos."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "PrSupxf85vwz",
+ "outputId": "f0d9c00b-308a-4e95-dc78-23a17aabd6cd"
+ },
+ "source": [
+ "df.query('tempo_24 == 1')['atv_moa'].value_counts()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "1 4718\n",
+ "0 3054\n",
+ "Name: atv_moa, dtype: int64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 204
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "QS1GlDC7mvRt"
+ },
+ "source": [
+ "##8. Os experimentos que foram tratados com droga, 75% ativa com a espera de 72 horas."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "M-MT-MY99zsC",
+ "outputId": "7008fb93-82b4-4f5a-caae-93c40e12a90d"
+ },
+ "source": [
+ "df.query('tempo_72 == 1 and com_droga == 1')['atv_moa'].value_counts(normalize=True)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "1 0.659053\n",
+ "0 0.340947\n",
+ "Name: atv_moa, dtype: float64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 275
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3y6aHhDPm2ul"
+ },
+ "source": [
+ "Apenas 65% ativa o MoA em 72 horas"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "rMYcVDMH-LCj"
+ },
+ "source": [
+ "#***Criação da Maquina Preditiva*** \n",
+ "Objetivo 1: Prever quando ativa pelo menos 1 MoA"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "8BXbiJhg5oMn"
+ },
+ "source": [
+ "Então agora vamos criar maquina preditiva do primeiro objetivo.\n",
+ "\n",
+ "Vamos escolher uma semente para que sempre quando tiver parametros de aleatoridade usaremos essa semente.\n",
+ "\n",
+ "Vamos fazer a nossa seleção das variáveis target e variáveis preditoras. \n",
+ "\n",
+ " \n",
+ "\n",
+ "1. `Variável target:` variavel que queremos prever.\n",
+ "2. `Variável preditora:` variável que irá nos auxilar na predição.\n",
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ "Vamos dividir nosso conjunto de dados em conjunto de treino e conjunto de teste. \n",
+ "\n",
+ "geralmente o conjunto de teste tem de tamanho em torno de 20% a 50% de todo o conjunto de dados. Isso depende muito do tamanho do conjunto de dados. \n",
+ "\n",
+ "Neste projeto escolhi deixar o com 30% para testar o nosso modelo."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "gWbmRV1p-nYx"
+ },
+ "source": [
+ "SEED = 10\n",
+ "\n",
+ "x = df.select_dtypes('float64','int64')\n",
+ "y = df['atv_moa']\n",
+ "\n",
+ "x_train, x_test, y_train, y_test = train_test_split(x, y, test_size = 0.3, random_state = SEED)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "mdacPyZ18axe"
+ },
+ "source": [
+ "Vamos testar em vários modelos preditivos de `Classificação`, com o objetivo de selecionar o melhor modelo possível para nossos dados."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "9YMB3vF__By_"
+ },
+ "source": [
+ "##Modelo Regressão Logistica antes do Balanceamento"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "njKMYKCd7PTx"
+ },
+ "source": [
+ "O primeiro será o modelo de regressão Logistica, para isso precisamos importar a biblioteca `sklearn.linear_model` e escolher o modelo de `LogisticRegression`.\n",
+ "Para ver a documentação da Biblioteca [clique aqui](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) \n",
+ "\n",
+ "
\n",
+ " \n",
+ " \n",
+ " Real \n",
+ " Predictions \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 8579 \n",
+ " 1 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 16718 \n",
+ " 0 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 17007 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 1901 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 13470 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " ... \n",
+ " ... \n",
+ " ... \n",
+ " \n",
+ " \n",
+ " 619 \n",
+ " 0 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 12019 \n",
+ " 1 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 16869 \n",
+ " 1 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ " 7536 \n",
+ " 1 \n",
+ " 0 \n",
+ " \n",
+ " \n",
+ " 1551 \n",
+ " 0 \n",
+ " 1 \n",
+ " \n",
+ " \n",
+ "
\n",
+ "
7145 rows × 2 columns
\n",
+ "
***Balanceando os Dados*** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "5Slq9I8_E6Q9"
+ },
+ "source": [
+ "Desbalanceamento ocorre quando temos um conjunto de dados que possui `muitos exemplos` de uma classe e `poucos exemplos` da outra classe. Como vimos na coluna `com_droga`.\n",
+ "\n",
+ "Quando isso ocorre, temos algumas técnicas para aplicar , por exemplo o Upsample e Downsample. "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "RIUKkolyEfg3"
+ },
+ "source": [
+ "***Upsample:***\n",
+ "\n",
+ "\"Upsampling é um técnica de processamento digital de sinais para aumentar artificialmente a taxa de amostragem em N vezes, inserindo um número N-1 de zeros entre as amostras originais do sinal, e passando o conjunto obtido por um filtro de reconstrução, que nada mais é que um filtro do tipo passa-baixas. Para mostrar essa técnica utilizaremos o programa anterior ligeiramente modificado. O script a seguir simula o upsampling da senoide ideal e realiza a filtragem de reconstrução.\"\n",
+ "\n",
+ "***Downsample:***\n",
+ "\n",
+ "\"Downsampling ou decimação é a técnica de redução da taxa de amostragem. Isso é feito simplesmente separando uma amostra a cada N. Há uma inevitável distorção do sinal.\"\n",
+ "\n",
+ "Fonte: [Clique Aqui](https://www.embarcados.com.br/oversampling-upsampling-downsampling-dsp/#:~:text=A%20sobre%2Damostragem%20(oversampling),artificial%20da%20taxa%20de%20amostragem.)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "t35_Sx_xEDku"
+ },
+ "source": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "8TKQFsLVGVuv"
+ },
+ "source": [
+ "Veja que está muito desbalanceado, isso faz com que a maquina preditiva aprenda muito quando é droga, mas não aprende quando não é droga, assim fazendo ela errar.\n",
+ "\n",
+ "Vamos resolver esse problema e vamos testar nossos modelos novamente."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "P48LXArDGkYt"
+ },
+ "source": [
+ "Escolhi a técnica de `Upsampling`, assim vamos aumentar o número de `0 ` que tem nessa coluna:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "-ssYCHhuA3cP"
+ },
+ "source": [
+ "#criando variáveis para a coluna\n",
+ "w= df.drop(['com_droga'],axis =1)\n",
+ "z = df['com_droga']"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "dHOxbE7FG3yt"
+ },
+ "source": [
+ "Vamos separar em variáveis qual tem mais e qual tem menos."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "wluSw4-PA3_j"
+ },
+ "source": [
+ "classe_com_mais = df[df.com_droga == 1]\n",
+ "classe_com_menos = df[df.com_droga == 0]"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "MJdctYyXA54s",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "83ac5903-43e2-4e97-9974-bea33771ca26"
+ },
+ "source": [
+ "#Classe antes do Upsampling\n",
+ "classe_com_mais.shape"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "(21948, 883)"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 231
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "c9BHRNb5A64q",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "7a3eb2ce-9f38-4e0d-d6bc-15d981a84764"
+ },
+ "source": [
+ "#Classe antes do Upsampling\n",
+ "classe_com_menos.shape"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "(1866, 883)"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 232
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "uoWLJrbHHMit"
+ },
+ "source": [
+ "Agora vamos aplicar a técnica UpSampling, vamos importar a biblioteca `sklearn.utils` e escolher o `resample`."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "pRqy-2LfBH6Y"
+ },
+ "source": [
+ "classe_com_menos_upsample = resample(classe_com_menos, replace = True,\n",
+ " n_samples = 21948,\n",
+ " random_state = SEED)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "J6hVOgH8Hltv"
+ },
+ "source": [
+ "Depois de feito a técnica, vamos concatenar os dados com o `.concat`."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "HAdpe69iBJtJ"
+ },
+ "source": [
+ "dados = pd.concat([classe_com_mais, classe_com_menos_upsample])"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "2QTR5Aq3IB8t"
+ },
+ "source": [
+ "Vamos verificar se foi feita a técnica? vamos utilizar o .`value_counts()`"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "8JiH7RQVBL8H",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "367ad045-b5b3-4df4-ea22-22e57f0aea21"
+ },
+ "source": [
+ "dados.com_droga.value_counts()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "1 21948\n",
+ "0 21948\n",
+ "Name: com_droga, dtype: int64"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 235
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "W_3Q45VPBNLc",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "fdb12d0c-b4e4-456a-c9b1-976d0d551332"
+ },
+ "source": [
+ "dados.info()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "\n",
+ "Int64Index: 43896 entries, 0 to 8203\n",
+ "Columns: 883 entries, id to atv_moa\n",
+ "dtypes: float64(872), int64(8), object(3)\n",
+ "memory usage: 296.1+ MB\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "uyUoOozFBRE4",
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 380
+ },
+ "outputId": "6c6a9b5c-ade1-4074-9934-c02a6651f825"
+ },
+ "source": [
+ "#Criando o Gráfico\n",
+ "sns.countplot(x='com_droga', data = dados, palette=\"crest\")\n",
+ "plt.title('Tipos de Tratamentos', fontdict={'fontsize':18, 'fontweight': 'bold'},)\n",
+ "plt.xlabel('Tratamento') \n",
+ "plt.ylabel('Quantidade') \n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "hoGQOlW4IPGs"
+ },
+ "source": [
+ "Pronto, agora a nossa coluna está balanceada, será que teremos uma melhora no nosso modelo? Vamos conferir!"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "T8Sjx5ecBTOu"
+ },
+ "source": [
+ "#**Criando Maquina Preditiva depois do Balanceamento**"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Va30I-CfIk3G"
+ },
+ "source": [
+ "Agora que temos a coluna balanceada, vamos dividir novamente nossos dados em treino e teste usando o `train_test_split` da biblioteca `sktlearn` ***Feature Selection (Seleção de Variáveis)*** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "_WTkfjPPETBQ"
+ },
+ "source": [
+ "Há boatos que para criar um modelo de aprendizado de máquina precisa de muita calma, paciencia e muito teste. Realmente, você precisa saber diversas técnicas, como elaborar, construir e refinar o seu modelo e testar! \n",
+ "\n",
+ "Uma dessas técnicas é a Seleção de Variáveis. Assim como o nome já diz, essa técnica seleciona as melhores variáveis para o modelo, ter variáveis de mais isso pode prejudicar a performance do algoritmo. \n",
+ "\n",
+ "Muitas features podem causar problemas como duração para treinamento do modelo ou dificuldades de colocar modelo de produção. Essa técnica ajuda a reduzir o overfitting e aumenta a accurácia do modelo e como já foi falado reduz o tempo de treino. \n",
+ "\n",
+ "Se quiser saber mais sobre essas técnicas em python [Clique Aqui](https://medium.com/data-hackers/como-selecionar-as-melhores-features-para-seu-modelo-de-machine-learning-faf74e357913)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "9sG8T_tzPKGN"
+ },
+ "source": [
+ "Para isso vamos utilizar o *SelectFromModel*, importanto a biblioteca `sklearn.feature_selection` e selecionar o `SelectFromModel`\n",
+ "\n",
+ "para acessar a documentação [clique aqui](https://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.SelectFromModel.html)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "a56qrp26Pv0p"
+ },
+ "source": [
+ "Vamos utilizar o modelo RandomForest para classificar as melhores variáveis, depois vamos treinar com RandomForest também para verificar as métricas."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "IQWvhPwpQMlV"
+ },
+ "source": [
+ "Para isso utilizei uma estrutura de repetição chamada For, para verificar em 10 e 10 quais as melhores variáveis.\n",
+ "\n",
+ "Deixei o código comentado pois ele demora muito para executar!"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "lkcLBgbwQbOu"
+ },
+ "source": [
+ "\n",
+ " '''k_vs_score = []\n",
+ "#for k in range(10, x_train.shape[1], 10):\n",
+ " model_rf = RandomForestClassifier(n_estimators=100,random_state=SEED, n_jobs=-1)\n",
+ " selector = SelectFromModel(model_rf, max_features=40, threshold=-np.inf)\n",
+ " selector.fit(x_train, y_train)\n",
+ "\n",
+ " x_train2 = selector.transform(x_train)\n",
+ " x_test2 = selector.transform(x_test)\n",
+ "\n",
+ " mdl = RandomForestClassifier(n_estimators=100,random_state=SEED, n_jobs=-1)\n",
+ " mdl.fit(x_train2, y_train) \n",
+ " p = mdl.predict(x_test2)\n",
+ "\n",
+ " score = mdl.score(x_test2, y_test)\n",
+ " print(\"k = {} - MAE = {}\".format(k, score))\n",
+ "\n",
+ "\n",
+ " mask = selector.get_support()\n",
+ " print(Xtrain.columns[mask])\n",
+ " k_vs_score.append(score)\n",
+ "'''"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "gKvgdmTgQpB2"
+ },
+ "source": [
+ "Após executar todo o código, foi observado que o melhor `valor` e onde o `score` está mais estável é quando o` K` é igual a `40`, `k` significa o número de variáveis,conforme imagem abaixo. "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "tt-B3hSdSZbd"
+ },
+ "source": [
+ ""
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "y62TjDIMSj__"
+ },
+ "source": [
+ "Vamos utilizar o `get_support() `para criar uma mascara e pegar as melhores variáveis que o `SelectFromModel` selecionou."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "CYy7iEETEMXC",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "da21f415-c36f-47f6-9444-35d1c01e1197"
+ },
+ "source": [
+ "model_rf = RandomForestClassifier(n_estimators=100,random_state=SEED, n_jobs=-1)\n",
+ "selector = SelectFromModel(model_rf, max_features=40, threshold=-np.inf)\n",
+ "selector.fit(x_train, y_train)\n",
+ "\n",
+ "x_train2 = selector.transform(x_train)\n",
+ "x_test2 = selector.transform(x_test)\n",
+ "\n",
+ "mdl = RandomForestClassifier(n_estimators=100,random_state=SEED, n_jobs=-1)\n",
+ "mdl.fit(x_train2, y_train) \n",
+ "p = mdl.predict(x_test2)\n",
+ "\n",
+ "score = mdl.score(x_test2, y_test)\n",
+ "print(\"k = {} - Score = {}\".format(40, score))\n",
+ "\n",
+ "\n",
+ "mask = selector.get_support()\n",
+ "print(x_train.columns[mask])"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "k = 40 - Score = 0.8124383020730503\n",
+ "Index(['g39', 'g50', 'g51', 'g57', 'g58', 'g68', 'g75', 'g100', 'g122', 'g138',\n",
+ " 'g175', 'g178', 'g206', 'g223', 'g230', 'g312', 'g317', 'g322', 'g328',\n",
+ " 'g352', 'g365', 'g367', 'g418', 'g445', 'g463', 'g524', 'g525', 'g529',\n",
+ " 'g600', 'g620', 'g635', 'g656', 'g689', 'g701', 'g738', 'g764', 'c13',\n",
+ " 'c65', 'c73', 'c98'],\n",
+ " dtype='object')\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "-PzoF46ZTJR2"
+ },
+ "source": [
+ "Esses são os nomes das colunas que o `SelectFromModel ` selecionou para ser as *melhores* `variáveis preditoras` do modelo."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "lQPIWjRWE2X_"
+ },
+ "source": [
+ "[['g39', 'g50', 'g51', 'g57', 'g58', 'g68', 'g75', 'g100', 'g122', 'g138',\n",
+ " 'g175', 'g178', 'g206', 'g223', 'g230', 'g312', 'g317', 'g322', 'g328',\n",
+ " 'g352', 'g365', 'g367', 'g418', 'g445', 'g463', 'g524', 'g525', 'g529',\n",
+ " 'g600', 'g620', 'g635', 'g656', 'g689', 'g701', 'g738', 'g764', 'c13',\n",
+ " 'c65', 'c73', 'c98']]"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "fi_SaOdCTism"
+ },
+ "source": [
+ "Agora vamos colocar no x as nossas novas variáveis preditoras"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "g_dscgJ_EVtH"
+ },
+ "source": [
+ "x = dados[['g39', 'g50', 'g51', 'g57', 'g58', 'g68', 'g75', 'g100', 'g122', 'g138',\n",
+ " 'g175', 'g178', 'g206', 'g223', 'g230', 'g312', 'g317', 'g322', 'g328',\n",
+ " 'g352', 'g365', 'g367', 'g418', 'g445', 'g463', 'g524', 'g525', 'g529',\n",
+ " 'g600', 'g620', 'g635', 'g656', 'g689', 'g701', 'g738', 'g764', 'c13',\n",
+ " 'c65', 'c73', 'c98']]\n",
+ "\n",
+ "y = dados['atv_moa']"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3OGyjUrQGTIi"
+ },
+ "source": [
+ "#**Criação de Maquina Preditiva após Balanceamento e Seleção de Variáveis**"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "RYlokH8cTy21"
+ },
+ "source": [
+ "Agora que já fizemos o balanceamento, a selecão de variáveis, vamos verificar se o modelo melhorou sua performance?\n",
+ "\n",
+ "***Criação da Maquina Preditiva*** \n",
+ "Objetivo 2: Prever se o experimento foi tratado com Droga ou Com controle."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "4Gt65VCKXKL1"
+ },
+ "source": [
+ "Agora vamos partir para nosso segundo objetivo! \n",
+ "\n",
+ "Prever se o experimento foi utilizado droga ou controle."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "-HQLpNOVXSRd"
+ },
+ "source": [
+ "Para isso vamos selecionar nosso` X e Y`\n",
+ "\n",
+ "Nosso `y `será a coluna `com_droga` e vamos selecionar todas as variáveis que são inteiro ou float para ser nosso `x`\n",
+ "\n",
+ "após isso vamos fazer a divisão da base de dados em teste e treino, deixando 30% para testar."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "-2HVsErfID0T"
+ },
+ "source": [
+ "X = dados.select_dtypes('float64','int64')\n",
+ "y = dados['com_droga']\n",
+ "\n",
+ "X_train, X_test, y_train, y_test = train_test_split(x, y, test_size = 0.3, random_state = SEED)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "6iB40sXrXw_9"
+ },
+ "source": [
+ "Vamos utilizar nossos famosos modelos que já foram para o paredão diversas vezes no último objetivo. Quem será que vai ganhar esse objetivo? Pra quem vai a sua torcida?"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "sDCNDLRlIRn7"
+ },
+ "source": [
+ "##Regressão Logistica"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "5027OuDcIWeJ"
+ },
+ "source": [
+ "model_rl.fit(X_train, y_train)\n",
+ "y_predict = model_rl.predict(X_test)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "rfcpIlh2I-ZV",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "0f2ca93c-c938-4d89-fb64-2cbf8ea61014"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.85 0.88 0.87 6539\n",
+ " 1 0.88 0.85 0.86 6630\n",
+ "\n",
+ " accuracy 0.86 13169\n",
+ " macro avg 0.87 0.86 0.86 13169\n",
+ "weighted avg 0.87 0.86 0.86 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Z3pdlB7WIqkB"
+ },
+ "source": [
+ "##XGBoost"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "isSL5qBNIp_C",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "16edc44a-eb67-4d89-9b9d-fef9e694119d"
+ },
+ "source": [
+ "#Treinando o modelo XGB\n",
+ "model_x.fit(X_train,y_train)\n",
+ "model_x.predict(X_test)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "array([1, 0, 1, ..., 0, 0, 0])"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 263
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "W6I0bkDrJFKb",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "1ba1f091-1726-49e3-f473-62914f995219"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.85 0.88 0.87 6539\n",
+ " 1 0.88 0.85 0.86 6630\n",
+ "\n",
+ " accuracy 0.86 13169\n",
+ " macro avg 0.87 0.86 0.86 13169\n",
+ "weighted avg 0.87 0.86 0.86 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "8CwYWFE7Ig_K"
+ },
+ "source": [
+ "## RandomForest"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "d3uxPy0_Icqb"
+ },
+ "source": [
+ "#Treinando o modelo\n",
+ "mdl.fit(X_train, y_train)\n",
+ "y_predict = mdl.predict(X_test)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "kPG5Pu2ZIn9Z",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "b2c72156-8cc9-4115-a46f-99e637fecb64"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.98 1.00 0.99 6539\n",
+ " 1 1.00 0.98 0.99 6630\n",
+ "\n",
+ " accuracy 0.99 13169\n",
+ " macro avg 0.99 0.99 0.99 13169\n",
+ "weighted avg 0.99 0.99 0.99 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "tCOjx9rzZFDg"
+ },
+ "source": [
+ "O resultado é ? \n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "opAJ39naYbEe"
+ },
+ "source": [
+ "Modelo |Resultado | \n",
+ "-------------------|------------------|\n",
+ "Regressão Logística | 86% \n",
+ "XGBoost | 86%\n",
+ "RandomForest | 99%"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "ID_nyzRlZiR8"
+ },
+ "source": [
+ "RandomForest ganhou novamente!!!!\n",
+ "\n",
+ "Desta vez com um alto índice de acerto 99%!\n",
+ "\n",
+ "***Conclusão*** "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "97bkoWY1e6vm"
+ },
+ "source": [
+ " Concluímos os objetivos que tinhamos traçados no começo deste projeto, pois criamos dois modelos que possuem um score de 81% e 99%, um para prever se o MoA vai ativar nos experimentos e o outro para prever se os experimentos foram utilizados droga ou controle. "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Z9WtY9oygrp1"
+ },
+ "source": [
+ "
Neste projetos aprendemos um pouco mais sobre essa area incrivel de Bioinformativa, a explorar os conjunto de dados, balancear os dados, selecionar as variáveis, escolher os diversos modelos de classificão para prever a nossa variável target. \n",
+ "\n",
+ "Mas não acaba por aqui! Ainda tem muitas coisas para fazer nesse projeto, como:\n",
+ "\n",
+ "\n",
+ "\n",
+ "* Balancear as outras colunas;\n",
+ "* Testar novos modelos de classificação;\n",
+ "* Tratar mais os dados;\n",
+ "* Normalizar os dados;\n",
+ "* Padronizar os dados;\n",
+ "* Prever outras variáveis;\n",
+ "\n",
+ "Obrigado por ter chegado até aqui!\n",
+ "\n",
+ "\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "LtchdLyne_QP"
+ },
+ "source": [
+ "#***FIM*** "
+ ]
+ }
+ ]
+}
\ No newline at end of file
![\"Open](\"https://colab.research.google.com/assets/colab-badge.svg\"){kind=icon}
![\"imersão](\"https://media.giphy.com/media/OWfzhRSn408151pwcU/source.gif\")
"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Gte_glj5jeHD"
+ },
+ "source": [
+ "![\"Paredawn\"](\"https://media1.tenor.com/images/70afa534ef1f780fadf92fa3ed56a805/tenor.gif?itemid=20867335\")
\n",
+ "\n",
+ "\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "v2p_CrKixE0k"
+ },
+ "source": [
+ "##Base Experimentos"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "2a2D7ZoOxl2T"
+ },
+ "source": [
+ "Para iniciar a nossa análise exploratória vamos começar entendendo a dimensão na nossa base utilizando o `.shape`:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "CAQj2A4exIJ9",
+ "outputId": "facbb5ae-68ec-42b3-e92d-903a0c983929"
+ },
+ "source": [
+ "df_experimentos.shape"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "(23814, 877)"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 144
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "fPh-ybHVx5u2"
+ },
+ "source": [
+ "Conforme observado acima, conseguimos perceber que temos `23.814 linhas` e `877 colunas` "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "8MtOef0Ux9Lq"
+ },
+ "source": [
+ "Para verificar as primeiras linhas do nosso DataFrame utilizaremos `.head()`"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 244
+ },
+ "id": "LW6K5VPIyJmL",
+ "outputId": "75447055-6851-42d9-ecd3-b5a7c17c70a9"
+ },
+ "source": [
+ "df_experimentos.head()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/html": [
+ "![\"Correlação](\"https://d33wubrfki0l68.cloudfront.net/e3b158e0207e1a80248bc0d7283f6eee638d5e72/4f6f0/blog/correlacao_files/figure-html/unnamed-chunk-4-1.png\")
"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "T2aSfzmP1WoC"
+ },
+ "source": [
+ "Vamos verificar qual a correlação entre as colunas dos genes: `g50` até `g100`e viabilidade celular: `c50` até `c99` tem entre elas:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "Kg-tZNw_Karl"
+ },
+ "source": [
+ "corr_g = df_experimentos.loc[:,'g50':'g100'].corr()"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "c09_FaYu1oKL",
+ "outputId": "e9a832d5-83df-4ad9-8ae6-ff42111eefad"
+ },
+ "source": [
+ "# Generate a mask for the upper triangle\n",
+ "mask = np.triu(np.ones_like(corr_g, dtype=bool))\n",
+ "\n",
+ "# Set up the matplotlib figure\n",
+ "f, ax = plt.subplots(figsize=(11, 9))\n",
+ "\n",
+ "# Generate a custom diverging colormap\n",
+ "cmap = sns.color_palette(\"crest\", as_cmap=True)\n",
+ "\n",
+ "# Draw the heatmap with the mask and correct aspect ratio\n",
+ "sns.heatmap(corr_g, mask=mask, cmap=cmap, center=0,\n",
+ " square=True, linewidths=.5, cbar_kws={\"shrink\": .5})\n",
+ "plt.title('Correlações entre g50 até g100',fontdict={'fontsize':18, 'fontweight': 'bold'})\n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "Zmpr2DANKfgl"
+ },
+ "source": [
+ "corr_c = df_experimentos.loc[:,'c49':'c99'].corr()"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "oWPoSM9q1wdr",
+ "outputId": "fe7e618f-97bb-43a7-94cf-a00d1653a0c0"
+ },
+ "source": [
+ "# Generate a mask for the upper triangle\n",
+ "mask = np.triu(np.ones_like(corr_c, dtype=bool))\n",
+ "\n",
+ "# Set up the matplotlib figure\n",
+ "f, ax = plt.subplots(figsize=(11, 9))\n",
+ "\n",
+ "# Generate a custom diverging colormap\n",
+ "cmap = sns.color_palette(\"crest\", as_cmap=True)\n",
+ "\n",
+ "# Draw the heatmap with the mask and correct aspect ratio\n",
+ "sns.heatmap(corr_c, mask=mask, cmap=cmap, center=0,\n",
+ " square=True, linewidths=.5, cbar_kws={\"shrink\": .5})\n",
+ "plt.title('Correlações entre c49 até c99',fontdict={'fontsize':18, 'fontweight': 'bold'})\n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "iVBORw0KGgoAAAANSUhEUgAAAk4AAAIbCAYAAAAdGhGxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdfVxUZf7/8dcIaoUKoeI6SpuaSnEXKHiXCuUtlJlpm5tWm6216ea3NcWw1MS7Sf221rqttrWubq2uu5Dld9XdLE0WUMsslLSyMnUNEckiRSHm9wcxP009c40yN+r7+Xj4eDDM4fp8zjWHuDrnzHtsTqfTiYiIiIi4Vc/fDYiIiIhcLLRwEhERETGkhZOIiIiIIS2cRERERAxp4SQiIiJiSAsnEREREUNaOMkF+/DDD5k4cSI333wzsbGxxMXFMXDgQGbMmMH+/fv93d45jRw5ko4dOzJy5Eiv1tmwYQPdunUjJiaGm2++mXvuuYf+/fvTp08fr9b1pZUrV7JkyRJ/t3FOJ0+e5NZbb6Vjx4507NiRzZs3u56rrq7mz3/+M4MGDSIuLo6EhATuv/9+3n///TrtwWSOZs6cSceOHRk6dChff/11ndQ13T9fzYPIxc6mHCe5EIsXL2b+/PkAXHnllURFRVFeXs5nn33G999/T5MmTVi8eDEJCQl+7vRMI0eOZMuWLSQnJ7Ns2TKv1Rk0aBDHjh0jIiKCgwcPUlJSwlVXXcXUqVNJT0/3Wl1fKS8vp1u3bjRv3py33nrL3+2c1fPPP8/vfvc71+OlS5fSpUsXAKZOncry5csBaN++PeXl5Rw8eJAGDRrwyiuvEBcXd8H13c2R0+kkKyuLFStW8PDDD/OrX/2K4ODgC64L5vvni3kQuRTojJOct02bNrkWTX379mXTpk0sX76c1atXs3r1atq3b88333zDuHHjOHHihJ+79Z/XX3+dN998k1dffZW3336bHTt2sGXLlkti0QTwzjvvcPLkSX+3cU6ffPIJixYtomPHjmc8t3//flasWAHULKRXr17Nm2++Sa9evTh58iTz5s2rkx7czVFZWRnXX389K1eu5Ne//nWdLZpM989X8yByKdDCSc7bokWLAGjRogXz5s2jcePGrufatm3Ls88+S/v27Rk0aBDl5eWu5/bv388TTzxBr169iImJoWvXrowZM4bCwsLTxq+9lJaZmcny5cvp1q0b48aNc/sc1Fw+fPjhh0lKSiIuLo4hQ4awdu1ao/1atWoVQ4YMITY2luTkZEaMGEF+fv4Z2+3evZtHH32ULl26EBMTw4ABA3jllVeoqqo6bTvT/TXtu6qqiiVLlnDHHXfQuXNnEhMTufPOO11/+NwxqXHq/BYVFXHPPfcQHx/PTTfdxJ/+9KfTtnvssccAOHDgwGmXPr35GpnOfXV1NU8++SQA48ePP2OcHTt2UHvS/c477wQgODjYtQ9bt26lrKzMshd3x4vVHNXOQ2ZmJs888wx33313nc6D6f7VxTyIXC60cJLzUl5eznvvvQfAgAEDuOKKK87Ypn379qxevZrHH3+cpk2bArBnzx7uvPNOsrOz+fbbb7n++uv5/vvvefPNNxk+fPhZFyiffvopM2fOxG6385Of/MTtc++++y4///nPefvtt7nyyiu57rrrKCoqYty4cfzzn/+03K8lS5YwceJEdu7cSfv27QkJCWHr1q08+OCDfPjhh67tCgsL+dnPfsa6deuoV68e119/PXv37mX69OlMnTrVtZ0n+2vat8PhYPbs2XzyySe0bduWtm3bsmvXLqZMmcILL7xguX+ezs1XX33FAw88QHl5OfXr16ekpIQ5c+awevVqAKKjo2nevDkADRs2pFu3bkRHR3v1NbKa+6effvq0bZctW8b27dt58MEHadeu3RljVVdXn7VGo0aNXM9//PHH5+zF5HixmiNvz4Pp/l3oPIhcVpwi56GoqMjZoUMHZ4cOHZzLli0z/rn777/f2aFDB2ePHj2cBw4ccDqdTmdZWZmzf//+zg4dOjj79u3rrK6udjqdTueIESNcNVatWnXaOFbPDR482NmhQwfnXXfd5Txx4oTT6XQ6//KXvzg7dOjg7NWrl/P7778/bYwRI0Y4nU6ns7q62tmrVy/n9ddf7/zNb37jdDqdzuPHjzu7dOni7NChg/Pxxx931fjZz37m7NChg7N///7Ob7/91ul0Op05OTmunoqKijzeX9O+b7zxRmeHDh2cr7/+uqufNWvWOG+99Vbn+PHjLeff07np0KGD829/+5vT6XQ6v/rqK1ftBx980DVmRkaGs0OHDs7U1NQ6f43Oxt3c79692+l0Op379+933njjjc6+ffs6KyoqnPv27XNtU1BQ4HQ6nc49e/a4vpeVleV0OmuOg8mTJ7u+/89//vOsfXhyvJxrjrw9D6b7dyHzIHK50RknOS/Hjh1zfR0aGmr0M2VlZa4zLHfeeSd2ux2AsLAwhg8fDsDevXv59NNPT/u5pk2bcuutt551zB8/V1xcTFFREQBpaWk0aNAAqLlBG2rOoOzateusY9lsNjZu3EhRUZHr3q0rrriCNm3auMYGKC0tdb3TaNCgQa7/K09PT+cPf/gDf/jDH2jSpIlH++tJ31deeSVQ8y6td955h/LycgYMGMAbb7xheS/K+cxNs2bNXJduWrRoQadOnQA8erdkXb5GJnNf+70pU6Zw7Ngxpk2bRsOGDc86Xtu2bV29LVu2jLS0NPr16+c6owac894k0+PlXHwxD6b7dyHzIHK50cJJzsup9zMdOXLE6Ge+/PJL130U11577WnPXXPNNa6vf/xH2W63U6/e2Q/VHz/31Vdfub6eNWuW6+3nnTt3dn3/v//97zl73LFjBw8//DDdunVz/ey2bdsAXL2f2l/tYgigfv36pKamkpqaSqtWrTzaX0/6Hj9+PDabjc2bN/PLX/6SpKQkhg4dytKlSy1vwj+fuYmMjDxtfmsvuXpys39dvkYmc2+323nttdfIzc3ltttuo3v37pb9zZ49m1GjRhEeHs6+ffu4+uqref75513Ph4SEnPNnTY6Xc/HFPHiyfxcyDyKXk7p564Zcdq699lrq169PZWXlGWeITvXZZ5/Rtm3bM77/43cNnfr4x4uk2jMsZ/Pj5079Y9WhQwfXH3qr2rUOHDjAL37xC7755hsaNGhAQkICV1xxBTt37uSbb745aw13fxzPVfPH++tJ33feeSexsbH8/e9/p6CggE8//ZTCwkIKCwvJzc1l8eLFZ+3hfOamfv36pz222Wzn2sVzqsvXyHTu//GPfwA17/zs1asXcPr9PuPGjeOnP/0pK1asoEGDBkycOJGJEye6nt+3b5/r61MXJqcyPV7OxRfzABjv3/nOg8jlRgsnOS8NGjSgV69erF+/ntWrVzNx4sTTzkIB7Nq1i8GDB3PDDTcwffp0IiMjXc/9eLG1Z88e19c//elPz7uvU//jfvfdd3PPPfcY/+y///1v1x+8xYsX061bNwBuu+220/4Qtm7d2vX1qX9YagMEq6urSUpK8mh/r7rqKo/67tChA5mZmUDNjfqvvvoq8+fPZ+PGjezevfusb72/kLmpSxfSh8nc1+YzAecMkSwrK3PNeXFxMbt27SImJsa1eKk9a3TVVVdx3XXXnXUM0+PlXHwxDzExMcb7d77zIHK50aU6OW9jxowhODiYY8eO8dhjj532duU9e/Ywbtw4nE4nBw4cIDIykvDwcLp27QrUnBE49Z6hV199FYCoqKgzLmt5IiIigqioKACys7M5fvw4UPOW74ceeohJkyZRWlp61p899UzX1VdfDdSkfte+m6j2j2GzZs2Ij48Hat6KfvToUQDWrVvHnDlzeOaZZ6isrPRof0373rNnD3fffTc333wzJSUlQM07n/r37+/q/VxnIC5kbqwEBQUBNYuR2jGtXEgfpnO/bNkydu/efdq/9evXu8ZZunQpb731FidOnCA9PZ3Ro0fz8ssvA3D8+HHX16fee/RjpsfLuebIF/Ngun8XMg8ilxudcZLzFh0dzZw5c3jiiSfYtGkTvXv3JioqiuPHj/PZZ59RVVVFkyZNePHFF103kE+ZMoWf//znlJSUMHDgQK677jr27NlDeXk5ISEhZGVlXXBfkyZN4sEHH2THjh3ccsstXHPNNRQVFXHixAluu+22s14SAejSpQtBQUF8//33/OIXv6Bly5Z8+umnPPjgg/zxj39k165dDBs2jN///vdkZmZy3333sX//fvr168c111zDjh07ABg8eLDrJmpP9tek76uvvpqTJ09y4MAB+vfvT1RUFE6n03UTcWJi4lnPNl3o3FipXegeO3aM9PR0EhISXDdLe6MPd3PvSUp9w4YNuf/++3n++ef54x//SH5+PiUlJRw6dIgWLVqcljv1Y54cL+eaI1/Mg8n+Xcg8iFxugqZNmzbN303Ixatjx46kpaVRWVnJ119/zd69eykrK6NVq1YMGjSIuXPnnpafEx4ezoABA/j22285ePAgX375JY0bN6ZPnz7MmzfP9X/gADk5ORw4cIBWrVoxZMiQ0+paPRcZGUnXrl356quvOHDgAAcPHqRVq1Y88MADTJgwwfV//z8eo1mzZlxzzTV8/PHHHDlyhPDwcGbPns0dd9xBYWEhBw8eJCgoiLvuuou2bdvSu3dvSkpK+OijjyguLubaa6/l0UcfZdy4ca6zEZ7sr0nfNpvNNd/FxcV88cUXHDlyhNatWzN06FCefvrpc76D7ELmptabb77Jrl27aNKkCffdd5/rGPjkk0/46quvOHHiBG3btqVfv3518hqdzU9+8hN69+7NoUOHOHjwIAcPHuSnP/0pv/71r0+b+x/75ptvWLp0KQB33HGH63JXUlISjRs35uDBg3zxxRcEBQXRr18//vd///eM3LBTeXK83HjjjWedI1/Mg+n+ne88iFxu9Fl1IhfojjvucCVkN2nSxN/tiIiIF+keJ5ELcOLECWJiYli/fj0jRozwdzsiIuJlWjiJXIDg4GAmT57MDTfcwKFDh/zdjoiIeJkWTiIXYN68eSQmJlJSUsKsWbP83Y6IiHiZ7nESERERMaQzTiIiIiKGfJLjVFLyreXzzZs3Nt4uUMfyR01fjuWPmur//MbyR03179+a+l0yH8sfNQO9f/GMzjiJiIiIGNLCSURERMSQ5aW66upq1qxZQ25uLqWlpTidTlq1akVqaiq9e/f2VY8iIiIiAcFy4TRt2jRatmzJ8OHDyc3Nxel0Eh8fT3Z2NgUFBWRkZPiqTxERERG/s1w4ff7550yfPh2AuLg47r//fsaMGcNNN93EsGHDfNKgiIiISKCwvMfJ6XSSm5vL0aNHWbVqFVdccQUAGzdu9ElzIiIiIoHE8ozT9OnTmTdvHnv37qVjx448/fTTAHz44Yc888wzPmlQREREJFB4nByen59Pt27dvNWPiIiISMCyPOP02muvnfbY6XTywgsv8MgjjwAwePBg73UmIiIiEmAsF04LFy4kLCzstOiBEydOsH//fo+KxHSZYPn8js1zgcBLaFVasH9rqv/zG8sfNdW/f2vqd8l8LH/UDPT+xTOWC6fVq1fz+9//nt27dzNp0iRatWrFpk2bGDt2rK/6ExEREQkYlgunhg0b8thjj/HZZ5+RlZXFjTfeSHV1ta96ExEREQkoRh+50qhRI3r37s1PfvITWrduzeLFizl06JC3exMREREJKEYLp4yMDEJDQxk8eDDPPvss7du3Z9KkSd7uTURERCSgGC2cKioqSEtLcz1OTU2lsrLSa02JiIiIBCLLe5xq2e12HA4HiYmJVFdXU1BQgN1u93ZvIiIiIgHFaOHkcDjIyckhLy+PoKAg4uPjSU9P93ZvIiIiIgHF4+RwERERkcuV0RmnCxXT1U0AZkFNAGZMt4nW2+U/E7ChZf6oebGHrql/74zlj5rq37819btkPpY/agZ6/+IZo5vDRUREREQLJxERERFj571wmjdvXl32ISIiIhLwLO9xOn78+Dmf2759e503IyIiIhLILBdOSUlJREREnPY9m82G0+mktLTUq42JiIiIBBrLhdPEiRMpLS3lscceO+O5kSNHeq0pERERkUBkeY/TvffeS5s2bTh27NgZz/Xo0cNrTYmIiIgEIrc3hw8ePJirrroKgKqqKg4cOEBVVRUPP/yw15sTERERCSSWyeEzZszgySefBCAvL4/JkyfTrFkzSktLefrpp+nZs6fPGhURERHxN8t7nHbv3u36euHChSxdupTIyEhKSkoYO3as8cIpNvlxy+cLt9REG5gkjMcmjrcea9t8QGnBdT2WP2qq//Mbyx811b9/a+p3yXwsf9QM9P7FM5aX6mw2m+vr0NBQIiMjAWjevDnBwT75tBYRERGRgGG5+vnkk08YN24cTqeTL7/8kjVr1jBw4EBefvllGjfWKlVEREQuL5YLpwULFgBQXFxMeXk5nTp1AuDkyZNMnTrV+92JiIiIBBDLS3XJyckkJyeTnZ1NaGioKwwzKiqKyZMn+6RBERERkUBh9Fl1FRUVpKWluR6npKRQWVnptaZEREREApHRHd52ux2Hw0FiYiLV1dUUFBRgt9u93ZuIiIhIQDFaODkcDnJycsjLyyMoKIj4+HjS09O93ZuIiIhIQLEMwBQRERGR/8/oHicRERERMbxUd6Gie02yfH7nO3Nqtuv9hPV2G2cT08VNuvjmuQDEdJtovV3+M0oL9mAsf9RU/+c3lj9qqn//1tTvkvlY/qgZ6P2LZ3TGSURERMSQFk4iIiIihiwv1ZWVlbFy5UpatGjB7bffzqJFi9i2bRtt2rRh9OjRhIeH+6pPEREREb+zPOM0ceJETp48yXvvvceYMWP49ttvGTNmDK1bt2biROt7iEREREQuNZZnnE6cOMHYsWNxOp0MGDCAhQsXAhAXF8e6det80qCIiIhIoLA841RVVcWBAwew2Ww8+eSTru/v2rVLH7kiIiIilx3LhdOECROYO7fm7f09e/YE4M033yQjI4OnnnrK+92JiIiIBBCPksOrqqooLi6mRYsWBAf7JAJKREREJGBYrn5mzJjhukSXl5fH5MmTadasGaWlpTz99NOus1DuxHR1E1pZMNd4u9hO4y23KXxvPoDRdqZ9XaxBaYEeuqb+vTOWP2qqf//W1O+S+Vj+qBno/YtnLBdOu3fvdn29cOFCli5dSmRkJCUlJYwdO9Z44SQiIiJyKbC8x8lms7m+Dg0NJTIyEoDmzZvrUp2IiIhcdixXP5988gnjxo3D6XTy5ZdfsmbNGgYOHMjLL79M48Y6vSciIiKXF8uF04IFCwAoLi6mvLycTp06AXDy5EmmTp3q/e5EREREAojlpbrk5GSSk5PJzs4mNDSUiIgIAKKiopg8ebJPGhQREREJFEYf8ltRUUFaWprrcUpKigIwRURE5LJjdIe33W7H4XCQmJhIdXU1BQUF2O12b/cmIiIiElCMAjCrqqrIycmhqKiIoKAgYmJiSE9Pp379+r7oUURERCQgeJQcLiIiInI580kYU0y3iZbP78h/xni72KTHLbcp3DoPgNhEN8nh2+Yb92WaVh5oCbOBnlar/r0zlj9qqn//1tTvkvlY/qgZ6P2LZ4xuDhcRERERLZxEREREjFkunDZu3Oj6+uuvvyYrK4uRI0eSlZXFkSNHvN6ciIiISCCxXDi99NJLrq+zsrJo0aIF06ZNo127dmRmZnq9OREREZFAYnxz+OHDh5k/v+Ym6Hbt2rFmzRqvNSUiIiISiCwXTmVlZa7LdQ0aNGDXrl1ERUWxb98+jh8/7pMGRURERAKF5cIpJiaGtWvXAtCsWTO+/vprAObOncuECRO8352IiIhIAPEoALOqqori4mJatGhBcLBPIqBEREREAobl6mfGjBk8+eSTAOTn55OZmUmzZs0oLS3l6aefpmfPnkZFons9Yfn8zndm12zX2812G2cT0yPDcpsd/3HUjNVzkvVYm+YYj1WXNRW65p+x/FFTAYznN5Y/aqp/87H8UVP9e7emeMZy4bR7927X17/73e9YunQpkZGRlJSUMHbsWOOFk4iIiMilwDKOwGazub4ODQ0lMjISgObNm+tSnYiIiFx2LFc/n3zyCePGjcPpdPLll1+yZs0aBg4cyMsvv0zjxjq9JyIiIpcXy4XTggULACguLqa8vJxOnToBcPLkSaZOner97kREREQCiOWluuTkZJKTk8nOziY0NJSIiAgAoqKimDx5sk8aFBEREQkURh/yW1FRQVpamutxSkoKlZWVXmtKREREJBAZ3eFtt9txOBwkJiZSXV1NQUEBdrvd272JiIiIBBSjhZPD4SAnJ4e8vDyCgoKIj48nPT3d272JiIiIBBSPksNFRERELmc+CWOKTX7c8vnCLfOMt4tNcrPN1h/G6uxmu3fnGSeCG9c06N8kXRwuj7Ra9e+dsfxRU/37t6Z+l8zH8kfNQO9fPGN0c7iIiIiIaOEkIiIiYsxy4TRlyhQKCwt91YuIiIhIQLO8x2n79u1UVVXx4osvMmLECJKTk33Vl4iIiEjAsVw4hYaGMmvWLD7//HOWLl3KzJkziYuLIyoqivDwcAYOHOirPkVERET8znLhZLPZAGjTpg1Tp06lsrKSrVu3UlhYyOeff66Fk4iIiFxWLBdOV1999WmP69evT/fu3enevbtXmxIREREJRB4FYFZVVVFcXEyLFi0IDvZJBJSIiIhIwLB8V92MGTNcX+fl5dG3b1/+53/+h379+rFp0yavNyciIiISSCxPG+3evdv19cKFC1m6dCmRkZGUlJQwduxYevbsaVTkhlsmWz5ftH6m8XbRKU9YbrNzw2wAonu72W7jbG5IzbSu9/asmrF6uRnrnR9qGvR2fZ8nLbf56M0ZxmNd7Gm16t87Y/mjpvr3b039LpmP5Y+agd6/eMbyjFPtzeFQ8w67yMhIAJo3b65LdSIiInLZsVz9fPLJJ4wbNw6n08mXX37JmjVrGDhwIC+//DKNG2uVKiIiIpcXy4XTggULACguLqa8vJxOnToBcPLkSaZOner97kREREQCiOWluuTkZJKTk8nOziY0NJSIiAgAoqKimDzZ+n4kERERkUuN0Yf8VlRUkJaW5nqckpJCZWWl15oSERERCURGd3jb7XYcDgeJiYlUV1dTUFCA3W73dm8iIiIiAcVo4eRwOMjJySEvL4+goCDi4+NJT0/3dm8iIiIiAcWj5HARERGRy5lPwphiuk+0fH5H3jMARN+UYbndzlwHscmPW25TuGUeALFJbrbbOo+Ybm76yq/pK6bLBOvtNs+tqZlsvV3hlrnGYZoxXd3ULJhrHOAZqKFrF3toXKCO5Y+a6t+/NfW7ZD6WP2oGev/iGaObw0VERERECycRERERY5aX6qqrq1mzZg25ubmUlpbidDpp1aoVqamp9O7d21c9ioiIiAQEy4XTtGnTaNmyJcOHDyc3Nxen00l8fDzZ2dkUFBSQkWF9T5KIiIjIpcRy4fT5558zffp0AOLi4rj//vsZM2YMN910E8OGDfNJgyIiIiKBwvIeJ6fTSW5uLkePHmXVqlVcccUVAGzcuNEnzYmIiIgEEsszTtOnT2fevHns3buXjh07Mm3aNAA+/PBDnnnmGV/0JyIiIhIwPArArKqqori4mBYtWhAc7JMIKBEREZGAYXmpbsaMGa6v8/Pz6du3L//zP/9Dv3792LRpk9ebExEREQkklqeNdu/e7fr6d7/7HUuXLiUyMpKSkhLGjh1Lz549jYrE9LB+992O/ziMtzNODu/sZrt3zZPD67KmaXK4SQq5aXK4SYq60nbrvqaSq89vLH/UVP/mY/mjpvr3bk3xjOUZJ5vN5vo6NDSUyMhIAJo3b65LdSIiInLZsVz9fPLJJ4wbNw6n08mXX37JmjVrGDhwIC+//DKNG2uVKiIiIpcXy4XTggULACguLqa8vJxOnToBcPLkSaZOner97kREREQCiOWluuTkZJKTk8nOziY0NJSIiAgAoqKimDx5sk8aFBEREQkURh/yW1FRQVpamutxSkoKlZWVXmtKREREJBAZ3eFtt9txOBwkJiZSXV1NQUEBdrvd272JiIiIBBSjhZPD4SAnJ4e8vDyCgoKIj48nPT3d272JiIiIBBSPksNFRERELmc+CWOK7jXJ8vmd78yp2a63m3DIjbOJ6WIdDLlj81wAo3DLmK5uxiowH8u45k1uQj5zfwgDNdjuhlusb9AvWj/TeKyY7ta9A+zIe0ahcRfBWP6oqf79W1O/S+Zj+aNmoPcvnjG6OVxEREREtHASERERMXbeC6d58+bVZR8iIiIiAc/yHqfjx4+f87nt27fXeTMiIiIigcxy4ZSUlORKC69ls9lwOp2UlpZ6tTERERGRQGO5cJo4cSKlpaU89thjZzw3cuRIrzUlIiIiEogsF0733nsvr732GseOHeOqq6467bkePXp4tTERERG5dMUmjvdo+8Jt873UiWfc3hw+ePDg0xZNR44cAeDhhx/2XlciIiIiAcgyOXzjxo2sX7+e6dOnk5+fT2ZmJiEhIRw7dowpU6aQkpLiw1ZFRETkUnGxnnGyvFT33HPPsWjRIgAWLlzI0qVLiYyMpKysjIceesh44WSc9m2Q5B2b9LjlNoVba2IS3L0ghdvmGyeC12XNaDcp3jt/SA6PTXZTc8s8olPcJK1vmP3DWNbzWrhlLrGd3B/Ahe/NN05bvxzSdgN1LH/UVP/+ranfJfOx/FEz0PsXz1heqquqqiIkJASAxo0b07p1awDCwsLQR9yJiIjI5cbyjNOoUaMYPHgwPXr0ICwsjEceeYSEhAQ2b97M0KFDfdWjiIiISECwXDgNGjSIXr16sWHDBvbv30+bNm1o1qwZSUlJpKam+qpHEREREbdmzZrFBx98gM1mIzMzk7i4ONdzb775Ji+88AINGjQgPT2dESNGuP2Zs3H7rrqwsDBWrVpF27ZtGT16NEOGDCEqKopJkyZd4O6JiIiI1I0tW7awd+9eVqxYwcyZM5k5c6bruerqarKysnjxxRd55ZVXePvtt/nqq68sf+ZcjD6rrqKigrS0NNfjlJQUKisrzwfMiTwAACAASURBVGO3REREROpefn4+ffr0AaBdu3YcPXqU8vJyAMrKymjSpAnh4eHUq1ePrl27kpeXZ/kz52J5qa6W3W7H4XCQmJhIdXU1BQUF2O32C9k/ERERkTpz+PBhoqOjXY/Dw8MpKSmhUaNGhIeH89133/HFF1/QqlUrNm/eTHJysuXPnIvRwsnhcJCTk0NeXh5BQUHEx8eTnp5+AbsnIiIi4j2nvvvfZrMxZ84cMjMzT0sJsPqZc7EMwBQRERHxhroOwHz++edp3rw5d999NwC33HILq1atOuvZo/nz5xMVFcVnn31m/DO1jO5xEhEREQlkPXr0YN26dQDs3LmTiIiI0xZADz74IKWlpRw7doy3336bbt26uf2ZszG6VHehTBO6jZLDDRK1AbdJ2IXvzSfGTYr3jtoU785uar5rXtM87dv9ft6Qmmm5TdHbs4zHcpeODjUJ6dG9rN9NufOdOcY1L/a03UAdyx811b9/a+p3yXwsf9QM9P4vFYmJiURHR3P33Xdjs9mYOnUq2dnZNG7cmL59+3LXXXfxwAMPYLPZGD16NOHh4YSHh5/xM+74ZOEkIiIichpb3Q/5+OOn/097VFSU6+t+/frRr18/tz/jji7ViYiIiBiyPONUVlbGypUradGiBbfffjuLFi1i27ZttGnTxnWaS0RERORyYXnGaeLEiZw8eZL33nuPMWPG8O233zJmzBhat27NxInW9y2JiIiIXGoszzidOHGCsWPH4nQ6GTBgAAsXLgQgLi7OdRe6iIiIyOXC8oxTVVUVBw4cwGaz8eSTT7q+v2vXLn3kioiIiFx2LBdOEyZMYO7cuQD07NkTqPl04YyMDJ566invdyciIiKXJpvNs38BwuPk8CNHjuimcBEREbkg7jISf6w2M9HfLO9x2rhxI+vXr2f69Onk5+eTmZlJo0aN+O6775gyZQopKSlGRYwDMA0CKU1CMk1rRvd0E+a4aY7xWHVe02A/r79lsuU2H62fadxXTA/ruQfY8R8H0W5eo50/hIYa1TQMIA3U0LhAHcsfNdW/f2te7AGM6t98LG/UFM9YLpyee+45Fi1aBMDChQtZunQpkZGRlJWV8dBDDxkvnEREREQuBW5vDg8JCQE47dOEw8LCjD5BWERERORSYnnGadSoUQwePJgePXoQFhbGI488QkJCAps3b2bo0KG+6lFEREQkIFgunAYNGkSvXr3YsGED+/fvp02bNjRr1oykpCRSU1N91aOIiIhIQHD7WXVhYWGsWrWKtm3bMnr0aIYMGUJUVBSTJlnf5CwiIiJyTjYP/wUIow/5raioIC0tzfU4JSVFAZgiIiJy2bG8VFfLbrfjcDhITEykurqagoIC7Ha7t3sTERERCShGAZhVVVXk5ORQVFREUFAQMTExpKenU79+fV/0KCIiIpeY2CQPAzC3XgQBmK6NgoMZNmyYt3sRERERCWhGC6cLFdPdTYp0Xk3ytkmqdkwXN4nam2uSw2OTrbcr3DLXOLnatKZRcnjvJyy32blxds1YdZkcbtC/u/mCmjkzTj43qGk8rwbHz6WeFnw5JVf7o6b6Nx/LHzXVv3drimeMbg4XERERER+dcRIRERE5TQBFDHjC8ozTxo0bXV9//fXXZGVlMXLkSLKysjhy5IjXmxMREREJJJYLp5deesn1dVZWFi1atGDatGm0a9eOzMxMrzcnIiIiEkiML9UdPnyY+fPnA9CuXTvWrFnjtaZEREREApHlwqmsrMx1ua5Bgwbs2rWLqKgo9u3bx/Hjx33SoIiIiEigsFw4xcTEsHbtWgCaNWvG119/DcDcuXOZMMH929dFRERELiVGyeGnOnLkCOHh4d7qR0RERC4DsckeJodvuQiSwzdu3Mj69euZPn06+fn5ZGZmEhISwrFjx5gyZQopKSlGRUzCKMEsQDK203jrsd6ruQ8rNtHNdtvmG9UDzGu6iY8v3DrPOHTTZKzoFDdhmhtqwjRN5t/kAC7cMs98zjq76f/deUbbgGmAp1n/F2vo3eUUwOiPmurffCx/1FT/3q3pPxdnHoHlwum5555j0aJFACxcuJClS5cSGRlJWVkZDz30kPHCSURERORSYBlHUFVVRUhICACNGzemdevWAISFheHhFT4RERGRi57lGadRo0YxePBgevToQVhYGI888ggJCQls3ryZoUOH+qpHERERkYBguXAaNGgQvXr1YsOGDezfv582bdrQrFkzkpKSSE1N9VWPIiIiIgHB7Yf8hoWFsWrVKtq2bcvo0aMZMmQIUVFRTJo0yRf9iYiIiAQMtwsngIqKCtLS0lyPU1JSqKys9FpTIiIicmlz2jz7FyiMPnLFbrfjcDhITEykurqagoIC7Ha7t3sTERERCShGAZhVVVXk5ORQVFREUFAQMTExpKenU79+fV/0KCIiIpcYdxl9P7Zj81wvdeIZozNOwcHBDBs2zNu9iIiIiAQ0o4XThYrp6ib5ucA8OdwkRRoMU7x7uEnx/s8PKd6GyecmCePRvaxvqt/5zhzjmjfcPNlym6K3ZgJ1m7wd7Sb5fGdt8rnBXLirWRuvH9PdzXGR94zbYwd+SJ43fC0DLS34ckqu9kdN9W8+lj9qqn/v1hTPGN0cLiIiIiI+OuMkIiIi4m2zZs3igw8+wGazkZmZSVxcnOu5V155hddff5169eoRExPD5MmTyc7OZsGCBVxzzTUAdO/enV/96leWNSwXTlOmTGHYsGHExsbWwe6IiIiI/KCOIwa2bNnC3r17WbFiBXv27CEzM5MVK1YAUF5ezksvvcS//vUvgoODeeCBB9i+fTsAaWlpZGRY34ZyKstLddu3b+evf/0rjz76KFu2bLmA3RERERHxnvz8fPr06QNAu3btOHr0KOXl5QDUr1+f+vXrc+zYMaqqqjh+/DihoaHnVcfyjFNoaCizZs3i888/Z+nSpcycOZO4uDiioqIIDw9n4MCB51VUREREpC4dPnyY6Oho1+Pw8HBKSkpo1KgRDRs2ZMyYMfTp04eGDRuSnp5OmzZteP/999myZQujRo2iqqqKjIwMbrjhBss6lgsnm63mPFqbNm2YOnUqlZWVbN26lcLCQj7//HMtnERERCQgnRpTWV5ezqJFi1i7di2NGjXivvvuY9euXcTHxxMeHk5KSgrvv/8+GRkZvPHGG5bjWi6crr766tMe169fn+7du9O9e/cL2BURERGRuhUREcHhw4ddjw8dOkTz5s0B2LNnD5GRkYSHhwPQuXNnduzYwdChQ2nXrh0ACQkJHDlyhO+//56goKBz1jFKDj/VkSNHXIVFREREzoe7jMcfq818PJdt27bx/PPP86c//YmdO3cyY8YM/vrXvwI1l/GGDx/OG2+8wRVXXMEvfvELxowZw/vvv0/Lli259dZb+fjjjxk/fvyFnXHauHEj69evZ/r06eTn55OZmUlISAjHjh1jypQppKSkeLTTIiIiIt6QmJhIdHQ0d999NzabjalTp5KdnU3jxo3p27cvo0aN4t577yUoKIiEhAQ6d+5M69atmTBhAsuXL6eqqoqZM2e6rWN5xunOO+9k0aJFNGvWjBEjRjB79mwiIyMpKyvjoYce4m9/+5vRzpimTUf3dJOqvWmOcfKzSVq2aQp5XSafm+yj6VjGyeEmieyGydsmYxnXNEgEB7Pjx10KPNQkwRu/lm5q7sh1BGzysD9qqn//1rzYk6vVv/lY3qjpLyZ/d05V+/fF3yzjCKqqqggJCQGgcePGtG7dGoCwsDA8vMInIiIictGzvFQ3atQoBg8eTI8ePQgLC+ORRx4hISGBzZs3M3ToUF/1KCIiIhIQLBdOgwYNolevXmzYsIH9+/fTpk0bmjVrRlJSEqmpqb7qUURERCQguP2Q37CwMFatWkXbtm0ZPXo0Q4YMISoqikmTrO/VEREREbnUuF04AVRUVJCWluZ6nJKSQmVlpdeaEhEREQlElpfqatntdhwOB4mJiVRXV1NQUIDdbvd2byIiIiIBxWjh5HA4yMnJIS8vj6CgIOLj40lPT/d2byIiIiIBxePkcBEREZELdbHmOBmdcbpQ0b3chD6+UxP6GN37CevtNs52G3S44z81YZom4YrGY5mGPhqMZ7KPYBbAeP0t1gGYH62vCcA0CpB0s01tTdMwU6O5MAwDNTl+3G1Tu51x6KZBb6bHjwIY1b+3a17sAYzq33wsb9QUzxjdHC4iIiIiWjiJiIiIGLO8VFddXc2aNWvIzc2ltLQUp9NJq1atSE1NpXfv3r7qUURERCQgWC6cpk2bRsuWLRk+fDi5ubk4nU7i4+PJzs6moKCAjAz398WIiIiInMHm7wbOj+XC6fPPP2f69OkAxMXFcf/99zNmzBhuuukmhg0b5pMGRURERAKF5T1OTqeT3Nxcjh49yqpVq7jiiisA2Lhxo0+aExEREQkklmecpk+fzrx589i7dy8dO3Zk2rRpAHz44Yc880xg5CmIiIiI+IrHAZhHjhwhPDzcW/2IiIjIZcBdrt6P1ebs+ZvlpbqNGzcyZcoUAPLz80lNTeXee+/l5ptvZsOGDb7oT0RERCRgWF6qe+6551i0aBEACxcuZOnSpURGRlJWVsZDDz1ESkqKURHzhGg3qdrvzCam6wTLbXYUzAUgpoub7TbPNU6RNk4ON+jthtRMy22K3p5lXNM0OdxkLmI7P265DUDhu/OMX8vYZOvxCrfMM+oLzJLD3dVz1TQ9fgzm3/S4MJkzJVf7t6b6Nx/LHzXVv3drimcsF05VVVWEhIQA0LhxY1q3bg1AWFgY+og7EREROW+2izOPwHLhNGrUKAYPHkyPHj0ICwvjkUceISEhgc2bNzN06FBf9SgiIiISECwXToMGDaJXr15s2LCB/fv306ZNG5o1a0ZSUhKpqam+6lFEREQkILj9rLqwsDBWrVpF27ZtGT16NEOGDCEqKopJk9x/Gr2IiIjIpcToQ34rKipIS0tzPU5JSaGystJrTYmIiIgEIstLdbXsdjsOh4PExESqq6spKCjAbrd7uzcRERGRgGK0cHI4HOTk5JCXl0dQUBDx8fGkp6d7uzcRERG5VF2cb6rzPDlcRERE5ELF3JTh0fY7ch1e6sQzRmecLpS7yamdDJOgwNhk6wDDwi0eBGCaBlsaBjUahSYazoVJOGddhmm628Z0O9ecmdTs4WYu/vPDXBjMWbTBL+DOXIf58WPQm2mAqrveduY6jI+LizW073IKkPRHzYs9gFH9m4/ljZriGaObw0VERERECycRERERY+e9cFqyZEkdtiEiIiIS+M574fTWW2/VZR8iIiIiAc/y5vBu3bqd9ftOp5Py8nKvNCQiIiKXvov1Lf2WC6c777yTVq1aMXz48DOeGzlypNeaEhEREQlElpfqxo8fzzfffMOxY8fOeK5jx45ea0pEREQkECkAU0RERHzOJH/vVDsDJADT8oxTYmIiWVlZlJaW+qofERERkYBleY9TdHQ0AwYMYPz48bRs2ZIhQ4aQkJBAcLBngePGCdEGScyxSY9bblO4dR4AsZ3dbPfuPOMUaeOayW622zKP6F5u0tHfmWM+VsoT1mNtmA2YJZ+7my+ombOYrm7GKjCfM+NEdpMUb9Pkc9OaBvtpnEJucFzXZTo9BF7a8eWUvO2Pmhd7crX6Nx/LGzXFM5YrIJvNRlJSEkuWLKGwsJCVK1fy1FNPERISQtOmTVm8eLGv+hQRERGxNGvWLD744ANsNhuZmZnExcW5nnvllVd4/fXXqVevHjExMUyePJnKykomTZrEf//7X4KCgpg9ezaRkZGWNSwXTqfe/hQbG0tsbCwAhw4doqSk5EL2TURERC5nNludDrdlyxb27t3LihUr2LNnD5mZmaxYsQKA8vJyXnrpJf71r38RHBzMAw88wPbt2/n8889p0qQJ8+fPJzc3l/nz5/Pb3/7Wso7lPU633347ULNQWr58uev7r732Gs2bN7/QfRQRERGpE/n5+fTp0weAdu3acfToUVfmZP369alfvz7Hjh2jqqqK48ePExoaSn5+Pn379gWge/fubNu2zW0dy4XT0KFDAcjIyKBJkyau77dv355Jk6zv1RERERHxlcOHD3P11Ve7HoeHh7uujjVs2JAxY8bQp08fUlNTiY+Pp02bNhw+fJjw8HAA6tWrh81m4+TJk5Z1jD5ypaKigrS0NNfj1NRUKisrPd4pEREREV849Xaj8vJyFi1axNq1a1m/fj0ffPABu3btsvyZczF6e5zdbsfhcJCYmEh1dTUFBQXY7XYP2hcRERHxnoiICA4fPux6fOjQIddtRXv27CEyMtJ1dqlz587s2LGDiIgISkpKiIqKorKyEqfTSYMGDSzrGJ1xcjgctG3blry8PLZu3Up8fDwzZsw4330TERERqVM9evRg3bp1AOzcuZOIiAgaNWoEQKtWrdizZw8VFRUA7Nixg2uvvZYePXqwdu1aAN5++226dOnito6Sw0VERMTn3OUa/lhtzqGVefPm8e6772Kz2Zg6dSpFRUU0btyYvn37snz5crKzswkKCiIhIYGJEyfy/fff8+STT/LFF1/QoEED5syZQ8uWLS1raOEkIiIiPueNhZMveBYBfp5M046NksM7jbfcpvC9+YBhcrVpCnaim5rbPKhpmKJuMlZ0bzfJ4RtnG49lnBxu+FqaJJ+bbAOGyeduUrxrxptrfPz4PDnc9Fg0Tbs3SM4P1ORkf9RU/+Zj+aOm+vduTfGM0T1OIiIiIuLmjNPJkyf597//TYsWLejcuTP/93//x3vvvUebNm246667aNiwoa/6FBEREfE7y4XTxIkTufLKKzl8+DBr167l66+/5pZbbuHDDz9k4sSJLFiwwFd9ioiIiPid5cKptLSUZcuWUVVVRd++fVm/fj316tVj4MCBjBgxwlc9ioiIiAQEy3ucKisr+e677wgODubRRx+lXr2azUtKSjhx4oRPGhQREZFLkc3Df4HBcuH0y1/+krFjxwJwxx13ALBp0yaGDRvGo48+6v3uRERERAKI5aW6W265hVtuueW073Xp0oX169cTFBTk1cZEREREAo1lAGZiYiJ33HEHjzzyCE2bNvVlXyIiInIJi+5lnUX4Yzvfme2lTjxjecYpOjqaAQMGMH78eFq2bMmQIUNISEggONiz3Mzr+z5p+fxH/6753Luofk9ZbrfrX1lcf8tk67HWzwTgBjfbFa2fyfVu6n30rywA45om+xk1YIrlNrvWTjceq8Ot0yy3+Xh1zfNG8+qmXm3Nupwz07Gi+rvpf12W232Emv00OS4Ao95M5hUwes1NxzL9XTLZznQsBUheXv0rQLLuawZ6/+IZyxWQzWYjKSmJJUuWUFhYyMqVK3nqqacICQmhadOmLF682Fd9ioiIiPid5cLp1Kt4sbGxxMbGAnDo0CFKSkq825mIiIhIgLF8V93tt98O1CyUli9f7vr+a6+9RvPmzb3bmYiIiFy6Ls40AuuF09ChQwHIyMigSZMmru+3b9+eSZM8+1RjERERkYud0Yf8VlRUkJaW5nqcmppKZWWl15oSERERCURGb4+z2+04HA4SExOprq6moKAAu93u7d5EREREAorRwsnhcJCTk0NeXh5BQUHEx8eTnp7u7d5EREREAoplAKaIiIiIN0T39jAAc+NFEIApIiIi4hUB9E45T/hk4RTTbaLl8zvyn6nZrrub7fKeIbbz45bbFL47D4DYTuOtt3tvPjFdJ1jXK5hbM1aSm5pbf6hp0FvMTRnWNXMdNWMluxlryzy3q/Xa1blJ/+7mCzycM4P5N9lHgJgubmpunut2m9rtjF9Lg+3q8rg22UegTo9Z07kwSYFX8rZ/a17sydXq33wsb9QUzxi9q05EREREtHASERERMWZ5qa6oqIgbbrgBgJMnT/LKK6/w8ccf0759e+655x4aNmzokyZFREREAoHlGac5c+a4vp45cyZ79+5lwIABlJaWkpmZ6fXmRERERAKJ8Yf87tmzh7/85S8A9O7dm5EjR3q3MxEREZEAY7lwqqioYM+ePTidTsLDw9m3bx+RkZF8++23fPfdd77qUURERC4xTtvFmUdguXC64oormDZtmuvx7t27iYyM5NFHH2X06NHe7k1EREQkoCg5XERERHzuhlTP7pUuenuWlzrxjOUZp8TERO644w7GjBlDeHj4eReJ6eEm9PE/NaGP0W7CIXfmOoxDB2OTrYMCC7fMJTrFTYDkhpoASZMAQ9PtTMIEjcfq+6T1WP+eARjOq5vXCGpeJ3cHeu2BbfI6mb6WJkGf7oJFoSZc1Dho0uCYNQ0gNZkzk9cIILrnJOvtNs0x3s40jDWq31OW2+36V5bx7/jFGjqoAEz/1lT/3q0pnrFcOEVHRzNgwAB+85vf0LJlS4YMGUJCQgLBwfqkFhEREbn8WK6AbDYbSUlJLFmyhMLCQlauXMlTTz1FSEgITZs2ZfHixb7qU0RERMTvjOMIYmNjiY2NBeDQoUOUlJR4tzMRERGRAGMZgHn77bcDNQul5cuXu77/2muv0bx5c+92JiIiIhJgLBdOQ4cOBSAjI4MmTZq4vt++fXsmTbK++VRERETkUmP0Ib8VFRWkpaW5HqemplJZWem1pkREREQCkdHb4+x2Ow6Hg8TERKqrqykoKMBut3u7NxEREZGAYhSAWVVVRU5ODkVFRQQFBRETE0N6ejr169f3RY8iIiJyibkkAzBdGwUHM2zYMG/3IiIiIhLQfJJkaZy8bZA+XJfJ4abJycZp013dbFcw1zht2mQs4xRvg/7d1autaZpcbdK/yWsEGL1O7o4L+CGt3KAvMOvf9Lg2SvE2TN42PmbrMLk9qr+b5PB1WcbzGt3LzVy8U3P8BFpas5LD/VtT/Xu3pt944UN+Z82axQcffIDNZiMzM5O4uDgAiouLefzxx13b7du3j/Hjx1NZWcmCBQu45pprAOjevTu/+tWvLGsoAlxEREQuelu2bGHv3r2sWLGCPXv2kJmZyYoVKwBo0aIFy5YtA2puPxo5ciQ333wz69atIy0tjYwM9x/bVcvoXXUiIiIigSw/P58+ffoA0K5dO44ePUp5efkZ2+Xk5NC/f39CQkLOq47lwmnx4sUUFxef18AiIiIivnL48GGuvvpq1+Pw8PCzfsrJypUrXTmVUHOmatSoUdx3330UFRW5rWN5qe4f//gH7777Ltdeey0jR44kMjLSk30QERER8YuzhQa8//77tG3blkaNGgEQHx9PeHg4KSkpvP/++2RkZPDGG29Yjmu5cIqIiGDx4sXk5+czc+ZMKioq6Nq1K1FRUYSHh7tuuhIRERHxp4iICA4fPux6fOjQoTM+Hm7Dhg1069bN9bhdu3a0a9cOgISEBI4cOcL3339PUFDQOetYXqqz/XDHe7du3fjDH/6Aw+GgadOmvPXWW7zwwgue75WIiIgIgM3Df2706NGDdevWAbBz504iIiJcZ5ZqFRYWEhUV5Xr84osvsnr1agA+/vhjwsPDLRdN4OaM049Pc7Vo0YJhw4Yp00lEREQCSmJiItHR0dx9993YbDamTp1KdnY2jRs3pm/fvgCUlJTQtGlT18/cdtttTJgwgeXLl1NVVcXMmTPd1rFMDnc6na6zTiIiIiJ15YZbJnu0fdF694saX7C8VNepUyeysrIoLS31VT8iIiIiAcvyUl10dDQDBgxg/PjxtGzZkiFDhpCQkEBwsGe5maZpwSYJy8Zp0wZp2aZj1WVN4+Rtg1Tq6BQ3KeQbfkghN0iIjk1+3HIbgMIt8zxIbrcez5OxTI4fd/MFNXNmWtMoeds0Ud4kEd90rDqcf9Pj+vp+1snhH/0ry7h/42PWICE9UJOf/VHzYk+uVv/mY3mjpnjGcgVks9lISkpiyZIlFBYWsnLlSp566ilCQkJo2rQpixcv9lWfIiIiIn5nfHN4bGwssbGxQM1b/M4WKiUiIiJyKbO8x+n2228HahZKy5cvd33/tddeOyMbQURERMSU08N/gcJy4VQbSZ6RkUGTJk1c32/fvj2TJlnfdyIiIiJyqTH6kN+KigrS0tJcj1NTU6msrPRaUyIiIiKByOjtcXa7HYfDQWJiItXV1RQUFGC3273dm4iIiEhAMVo4ORwOcnJyyMvLIygoiPj4eNLT073dm4iIiEhAsUwOFxEREfGG6z1MDv8oQJLDPUuyPE9+CR3s6ma7Ag/CKOswnDC6l5sAwHdmG9d0d9DVHmQmYYKmAZLGc2byWpoeFwZhoO5eb6h5zY3DTE36N5hXMAt2NdlHgGg3NXf+UNOof4PfETAMwKzDMFPT7UyPRQVgBn4Ao/o3H8sbNcUzPlk4iYiIiJzmIv0sXKN31YmIiIiIwRmn7du3k5ubS2lpKU6nk1atWpGamsp1113ni/5EREREAoblGadnn32Wv//977Rr146qqioaNWpEkyZNmDJlCkuWLPFRiyIiIiKBwXLhtG3bNmbMmMHAgQPJysrio48+4mc/+xnLli3j9ddf91WPIiIiIgHBcuF08uRJPvvsMwDeffddvv/+ewA+/fRT73cmIiIiEmAs73HKyMjgscce4/Dhw7Ru3ZqZM2ve3v7nP/+ZJ5980icNioiIyCXo4nxTnQIwRURExPeu7+vZCZiP/j3DS514xvJSXWJiIjNmzODIkSO+6kdEREQkYFleqouOjqZ///785je/oWXLlgwZMoSEhASCgz3LzTRNRTZJYo7t/LjlNoXvzgMwSog2TU42Tps2GM80bdqk5g2pmZbbFL09q6YvgxRyd/Vqa5omRBvNv2kiu8HxY9p/bJKb42drzfFTpyn2dZB8Xnss1uX8m86Fu/8r/OjfM4znwjj5vIeb35P/OOo0eV7J4eZj+aOm+vduTfGM5QrIZrORlJTEkiVLKCwsZOXKlTz11FOEhITQtGlTFi9e7Ks+RURERPzOcuF06u1PsbGxxMbGJv4GogAAIABJREFUAnDo0CFKSkq825mIiIhIgLG8x+n2228HahZKy5cvd33/tddeo3nz5t7tTERERCTAWC6chg4dCtTEEjRp0sT1/fbt2zNpkvUnk4uIiIick83m2b8AYfQhvxUVFaSlpbkep6amUllZ6bWmRERERAKR0dvj7HY7DoeDxMREqqurKSgowG63e7s3ERERkYBitHByOBzk5OSQl5dHUFAQ8fHxpKene7s3ERERkYCi5HARERHxuev7PeXR9h/9K8tLnXjGsyTL82QaIGkSyGccgJk43nq7bfPNwyhNa3ZyU/O9+UT3sr6pfuc7c2rGMug/OuUJ67E2zAbqJswRakIMjeesDl9Lo2BRN/sINftp/Foa9G8a7GoS+mgcEmsQDAlmoaemoZUmAZixyW7ma4v5awmGx6zpXNTB8VMbpnmxhiYGegCj+jcfyxs1xTNGN4eLiIiIiI/OOImIiIic6mK9T+i8zzgtWbKkDtsQERERCXznvXB666236rIPERERkYBneamuW7duZ/2+0+mkvLzcKw2JiIiIBCrLhdOdd95Jq1atGD58+BnPjRw50mtNiYiIiAQiy0t148eP55tvvuHYsWNnPNexY0evNSUiIiISiDwOwHQ6ndgC6MP2RERE5OIT1d+zAMxd6y6CAMzc3FxmzpxJeHg4GRkZPP300xw6dIiQkBCmT59OcnKyr/oUERERsTRr1iw++OADbDYbmZmZxMXFAVBcXMzjj///oN59+/Yxfvx4BgwYwKRJk/jvf/9LUFAQs2fPJjIy0rKG5cJp4cKF/PnPf+bo0aOMHDmSJUuWEBUVxYEDB5gwYQKvvvqq0Y6YpvKapCebJGqDYXK4YYqxSSI4mCWMm6RIA0ZJzKbJ4SbJ7e6SsuGHtGzD19JkLoyTww1qmvZvfPwYzJlxCrZBwrVx8rZh2rfJeKbH9Q23TLbcrmj9TPP+Tf9bYJIcbnos1sHxX7j1h0R5w/0MtLTpQE+uVv/mY3mj5qViy5Yt7N27lxUrVrBnzx4yMzNZsWIFAC1atGDZsmUAVFVVMXLkSG6++WZWr15NkyZNmD9/Prm5ucyfP5/f/va3lnUs73GqX78+ERERtG/fniZNmhAVFQVAq1atCAoKqov9FBEREblg+fn59OnTB4B27dpx9OjRsyYA5OTk0L9/f0JCQsjPz6dv374AdO/enW3btrmtY3nGKTQ0lGeffZaysjKuueYapkyZQs+ePdm+fTtNmzY9n/0SERERqXOHDx8mOjra9Tg8PJySkhIaNWp02nYrV67k5Zdfdv1MeHg4APXq1cNms3Hy5EkaNGhwzjqWCyeHw0FOTg7XXXcd5eXlhISE8J///IeWLVsyYsSI8945EREREW8623vf3n//fdq2bXvGYsrqZ37M8lLdVVddxT333EN2djahoaEMGjSIadOm0bFjRyZPtr7vQURERMRXIiIiOHz4sOvxoUOHaN68+WnbbNiw4bRw74iICEpKSgCorKzE6XRanm0Cw49cqaioIC0tzfU4JSWFyspKkx8VEREROZPN5tk/N3r06MG6desA2LlzJxEREWecWSosLHTdr137M2vXrgXg7bffpkuXLm7rWF6qq2W323E4HCQmJlJdXU1BQQF2u93kR0VERES8LjExkejoaO6++25sNhtTp04lOzubxo0bu24ALykpOe0e7bS0NPLy8hg+fDgNGjRgzpw5busYLZxq73XKy8sjKCiI+Ph40tPTz3PXREREROreqVlNwGlnlwDeeOON0x7XZjd5wuPkcBEREZELFTVgikfb71o73UudeMbojNOFMgm2BIzCIU0D6EzC/eoyTBAwCtEzDcA06c04ALMOwihrtzMODTUJIDWcf6MwRDdj1Y5nGrpp8loa928S7OqPAEzD0Mfr3QRgfrR+pnmYqUEYKBiGnprOmUFN49fS9Pg36E0BjP4Zyx81A71/8cz/a+/+o6Kq8/+BP8cBTWFlA4MYg9UgRPlhi6VpgaRYCmWbHd0lgdOuZz2WpkdTEVB+iKijppvSsazOapTbD5XNPtsvo1QUUdIw8wflj3UJN1RE1IwGnfv9A5lvbs69r7nOjBM+H385eHvd17zmzvXd3MtzRDeHExEREREXTkRERERiqpfqLBYLNm3ahKCgINxzzz3417/+hd27d6Nnz54YM2YMOnXq5K4+iYiIiG441YXTzJkz0blzZ5w+fRofffQRzp49i6FDh+Krr77CzJkz8cILL7irTyIiIqIbTnXh1NDQgJKSEly6dAnDhg1DWVkZOnTogBEjRvArV4iIiOimo3qPU0tLC3744Qd4eXlh8uTJ6NChdfNTp07hp59+ckuDRERERJ5CdeH017/+FZMmTQIAPP744wCA8vJyjB49GpMnT3Z9d0REREQeRPVS3dChQzF06NCrfnbvvfeirKwMRqPRpY0REREReRrV5PBt27ahqKgI/v7+yMzMREFBAU6ePAkfHx/MnTsX/fv3d2evRERE1E70GpHn0PY1Hxa4qBPHqH7i9OKLL2LNmjVoampCeno6Vq9ejcjISNTV1WHGjBlYu3ataCfOTNuVph2Lkp+FiebitGlJWrZGcvjXV5LDJfuMGqyRHL6lNTlclLyt8RoBra+TNPlZMn9npmVrPUfgStr0DTh+JK+5MxO1AdnMxMnhSbNVtzv46TzE9Fff375dV5K3hce/ZLbimTnh+Gmbq+Q9DsjOZdJ93gzJ1exfXssV+yTHqC6cvL29ERgYiMDAQHTt2tX2ZXndu3fnpToiIiK66agunPz8/LBs2TI0NjYiNDQUubm5iI+PR3V1NQICAtzVIxEREZFHUF04mc1mlJaWIjw8HBcuXICPjw+2b9+O4OBg5jgRERHRTUc1jqBLly4YO3YsNmzYAD8/P4wcORL5+fno1asXcnLUvzGdiIiIqL0Rfclvc3MzkpOTbY8TExPR0tLisqaIiIiIPJHqpbo2JpMJZrMZcXFxsFqtqKyshMlkcnVvRERE1F4ZbnQD+ogWTm33OlVUVMBoNKJv375ISUlxdW9EREREHkU1AJOIiIjIFXolOxiA+YFnBGCK7nEiIiIiIuGluuvlzITlmH4ayb27ryT3CrZzZoq0dDtnph1HJWgkh29tTQ4XJbJrzAu4MjNh2rdo/tLkcEHCslZydWs9eVq2JAnbqf07sS9AljwvTcTvM1T9N2gPlBWJa4mTwyUzk+5TcP6Rzl/yvgSE8xeeyyTfvPBrT65m//JartgnOYafOBEREREJqX7idODAAfTp0wcAYLFY8Oabb+Kbb77BXXfdhbFjx6JTp05uaZKIiIjIE6h+4rRw4ULbn4uKinD8+HEMHz4cDQ0NyM7OdnlzRERERJ5E9ROnn//C3ZEjR/DGG28AAAYPHoz09HTXdkZERETkYVQXTs3NzThy5AgURYG/vz9qa2sREhKC8+fP44cffnBXj0REREQeQXXhdMsttyA/P9/2uKamBiEhIZg8eTLGjx/v6t6IiIiIPIrqwqmkpOQXP1MUBX//+99d1hARERGRp1JNDt++fTvmzZsHf39/ZGZmoqCgACdPnoSPjw/mzp2L/v37u7NXIiIiaid+rcnhqp84FRcXY82aNWhqakJ6ejpWr16NyMhI1NXVYcaMGVi7dq1oJ84MvZNsA8hC76I0+trvQF8AROGW0tBKSehdb41gwoNlRQAgep5a8wKuzCx+lnqt8tbfxJTMX3pcSPrX2qZtO2kYq2hmwuNCVMuJswCc+16KfHiO6naHPi4UB7v2eVD9t3EPfD4fAETHmfi8IthOHEwrDfOVnAuc+VoK+/fUAMZfe4Dkr71/cozqwsnb2xuBgYEIDAxE165dERkZCQDo3r07jEajWxokIiKi9kcxGG50C7qoLpz8/PywbNkyNDY2IjQ0FLm5uYiPj0d1dTUCAgLc1SMRERGRR1BdOJnNZpSWliI8PBwXLlyAj48Ptm/fjuDgYKSlpbmrRyIiIiKPoJoc3qVLF4wdOxYbNmyAn58fRo4cifz8fPTq1Qs5Oer31xARERG1N6Iv+W1ubkZycrLtcWJiIlpaWlzWFBEREZEnUr1U18ZkMsFsNiMuLg5WqxWVlZUwmUyu7o2IiIjIo4gWTm33OlVUVMBoNKJv375ISUlxdW9ERETUXv06f6lOPQCTiIiIyBUiHsl3aPtv/k97+/nz52Pv3r0wGAzIzs5GbGys7e/++9//Ytq0aWhpaUGfPn0wd+5c7Ny5E1OmTMFdd93V2lNEBObMUc+uE33iREREROTJdu3ahePHj+Ptt9/GkSNHkJ2djbffftv29wsXLsRf/vIXDBs2DAUFBThx4gQAoH///li+fLl4P25ZOEUN1kjL3tKalh2VqLHd5gXihFxnJg9LE36jEjT2uXWhvJag/z4ayeEH2pLDBbW0Es2B1lRzcf+StG/B6w3I5qqVSA20plLLX0vthHdJX4AsLVuaqC0+Zp14LEYOz1Xd7tBHc8WJ8r2TZqtud/DTeQBk5wJpCn+fIRrvk8+K5MeFE49Z6cwktaTHhWQWTN52/j49vf/2YseOHUhKSgIAhIWFoampCRcuXICvry+sVit2796NpUuXAgDy8lq/7qW2ttbh/fATJyIiInI7yaU3R5w+fRpRUVG2x/7+/jh16hR8fX1x5swZ+Pj4YMGCBdi/fz/uuecePPfccwCAw4cPY8KECWhqasKkSZNw//33q+6HCyciIiJqd35+C7eiKKivr0dGRga6d++O8ePHY/PmzejduzcmTZqEESNGoLa2FhkZGfjkk0/QsWNHu3VVc5xWrVqF+vp65z0LIiIiIhcIDAzE6dOnbY9PnjyJ2267DQBw6623wmQyITQ0FEajEQMHDsS3336LoKAgJCcnw2AwIDQ0FN26ddNc96gunNavX485c+Zg/vz5uq4DEhEREbnD/fffj48//hgAsH//fgQGBsLX1xcA4OXlhZCQEPz73/+2/X3Pnj2xceNGvPbaawCAU6dOoaGhAUFBQar7Ub1UFxgYiFWrVmHHjh0oKipCc3Mz7rvvPkRGRsLf3/+qX/MjIiIiulHi4uIQFRWFP/3pTzAYDMjLy8OGDRvwm9/8BsOGDUN2djZmzZoFRVEQERGBIUOG4OLFi5g+fTrKysrQ0tKC/Px81ct0gMbCyWBoTacaOHAgBg4ciPr6emzduhWfffYZTp06hZUrVzrvGRMRERFdh+nTp1/1ODIy0vbn3/3ud/jHP/5x1d/7+vripZdecmgfqgun/83GDAoKwujRozF69GiHdkJERETUHqgmh1utVnTocPVtUIqi2D6JIiIiIrqZqH7itGPHDsybNw/+/v7IzMxEQUEBTp48CR8fHxQUFGDAgAGinYiD3gQhdNEDZ6pu8/WORQAg2k5c674Z6ttVLhZvJw3aix6k0VvFInFoojPmCjg4swEas9jpwCwE4X7SAE/JXAHh8SOtJelf+B6RBsBKjkXpa9n7IfWvHzj4SaG4liSAERAG2Ar36cwwUPFrLnidnHouE7zfAGkYrvo2QOvMGCB5Y2q5Yp/kGNWFU3FxMdasWYOmpiakp6dj9erViIyMRF1dHWbMmIG1a9e6q08iIiKiG0514eTt7Y3AwEAEBgaia9eutpusunfvDqPR6JYGiYiIiDyF6sLJz88Py5YtQ2NjI0JDQ5Gbm4v4+HhUV1cjICDAXT0SEREReQTVhZPZbEZpaSnCw8Nx4cIF+Pj4YPv27QgODkZaWpq7eiQiIiLyCKrJ4V26dMHYsWOxYcMG+Pn5YeTIkcjPz0evXr2Qk6N+kycRERFRe6O6cGrT3NyM5ORk2+PExES0tLS4rCkiIiIiT6R6qa6NyWSC2WxGXFwcrFYrKisrYTKZXN0bERERkUcRLZza7nWqqKiA0WhE3759kZKS4ureiIiIiDyKanI4EREREf1/ok+crpdWqvP+rQvE292Q5HDpdoJUYWcmREuTw0Up0hrJw0Br+rAz09YlidqALPlca66AY2nTkt6kKdKiFGzhLJyZ3C5NxBclhwtruTPt3pZWLtin+LUUpJADsuMnpr/6zPbtkqd9i5PDJedYjd6BK8e/8Ji9GZK3f+39k2NEN4cTERERERdORERERGKal+qqq6uxbds2NDQ0QFEUdO/eHQ8++CDCw8Pd0R8RERGRx1D9xGnZsmVYt24dwsLCcOnSJfj6+qJr167Izc3F6tWr3dQiERERkWdQXTjt2bMH8+bNw4gRI1BYWIiDBw/ij3/8I0pKSrBx40Z39UhERETkEVQXThaLBUePHgUAfPHFF7h8+TIA4PDhw67vjIiIiMjDqN7jlJmZialTp+LUqVMICQlBUVERAGDNmjWYPXu2WxokIiIi8hQOB2AqigKDweCqfoiIiIg8luqlum3btmHEiBEYO3YsvvrqKzzxxBNISEjA8OHDsWvXLnf1SEREROQRVD9xSk1NxQsvvICmpiakp6dj9erViIyMRF1dHWbMmIG1a9eKdiJOOxakZcf0e051m327nwcAxMRpbLfneXHarjP3KUlEBiBKFY5K1EgB3rzgSq3pGrWWaD5HoPV5ShOiJTOT9AUIU+CFyecx92js84vWfUp6k85C0r80RVp8zAqep/S47j00R3W7g2VF4rlK3uOAdGbC40eQCi4+roX7jLlXY7uqJfKZCV5zZ87f6Sn8gu1+7cnbv/b+yTGq9zh5e3sjMDAQgYGB6Nq1KyIjIwEA3bt3h9FodEuDRERERJ5CdeHk5+eHZcuWobGxEaGhocjNzUV8fDyqq6sREBDgrh6JiIiIPILqwslsNqO0tBTh4eG4cOECfHx8sH37dgQHByMtLc1dPRIRERF5BNWbw7t06YKxY8diw4YN8PPzw8iRI5Gfn49evXohJ0f9vgciIiKi9kb0Jb/Nzc1ITk62PU5MTERLS4vLmiIiIiLyRJpf8gsAJpMJZrMZcXFxsFqtqKyshMlkcnVvRERERB5FtHBqu9epoqICRqMRffv2RUpKiqt7IyIiIvIoDieHExEREd2sRJ84XS9x6J0kAFMaYCjYThx6JwizA4ShlQnq4XL7ty4EAFFv0gBMUYCexnMEWp+nNPROMgvpaykJDRUHYApfS1EAo7SWEwMYb8T8eyepfy/lwU/niYM5nRmA6cx9igNInXjOEO9TciwKw0xFAZiD1c8rALB/ywLnnteFs/DUAEkGYN5cRDeHExEREREXTkRERERiuhdOS5YscWYfRERERB5P9R6nH3/80e7fVVdXO70ZIiIiIk+munC69957ERgYeNXPDAYDFEVBQ0ODSxsjIiIi8jSqC6eZM2eioaEBU6dO/cXfpaenu6wpIiIiIk+keo9TRkYGevbsiYsXL/7i7+6//36XNUVERETkiUQBmIqioLGxEYqiICAgwB19EREREXkc1Ut1x44dg9lsRl1dHb777juEhYWhqakJUVFRyMrKQlBQkLv6JCIiIrrhVD9xysjIQFFREUJCQnD06FGUlJQgLy8PW7duxSuvvIKSkhLRTpyZiuzM5N6oeI0U7/LWFG9pqq0kiVkrlXf/lgVX9qlRa9cS9Bmao7rNgbIiAMLkYY3n2LpPefK5ZBZOTd7WeI6AY2nfouNHkGgOQHScid8jGvv8+so+JfWkydu9H5qjut3BTwrlKd7S/iXnAmktyfEjTMGWvC8BWVq/dGaS8480UV5yLPZ5MFt1GwA48Pl8px7/4rkKXicmhzu+T3KM6j1OFosFISEhAIAePXqgpqYGAJCQkIDm5mbXd0dERETkQVQv1UVERGDatGmIjY1FeXk5BgwYAADIzs5GeHi4WxokIiIi8hSqC6eCggKUlZWhpqYGSUlJSE1NBQCYTCaMGTPGLQ0SEREReQrVS3UGgwFJSUmoqqqCn5+f7efR0dGYNUv9ujURERFReyP6rjqLxYLk5GTb48TERLS0tLisKSIiIiJPpHqprk1wcDDMZjPi4uJgtVpRWVkJk8nk6t6IiIiIPIpo4WQ2m1FaWoqKigoYjUb07dsXKSkpru6NiIiIyKOIksOJiIiISPiJ0/WShss5MyhNsp08gM55+5SGbkoCDLWC6g58Pr+1r37Pqfe1+3nNwEqgNbRSGrQnmYU0WFRy/EgDPOWvpSB0UBpa6YT+bbNw5vyFYaCSAEzpXMXHvxMDJEUBpNIAT/H8tV9PSUisdJ/i+QtCK7XmBVyZmfC8LupfGoYrOS403pdA63uTAZhXb0dyopvDiYiIiIgLJyIiIiIx1Ut1jY2NePfddxEUFITHHnsML7/8Mvbs2YOePXti/Pjx8Pf3d1efRERERDec6idOM2fOhMViwe7duzFx4kScP38eEydOxB133IGZM7W/VJWIiIioPVH9xOmnn37CpEmToCgKhg8fjhdffBEAEBsbi48//tgtDRIRERF5CtVPnC5duoS6ujoYDAbMnj3b9vNDhw4xOZyIiIhuOqoLpxkzZmDRotZfr42PjwcAfPrpp8jMzMScOeq/okxERETU3ogCMBVFQWNjIxRFQUBAgDv6IiIiIvI4qvc4HTt2DGazGXV1dfjuu+8QFhaGc+fOoU+fPsjKykJQUJC7+iQiIiK64VQ/ccrIyEBRURFCQkJw9OhRlJSUIC8vD1u3bsUrr7yCkpIS0U7ECcuC7SQp2IAsLVval3ifcRrb7Xke0RrJvV9vM4v3GTU4S3Wb/VsWiPvS2qZtO3FysiStXDpXSQqzNDn8Bhw/0YM0ZlaxSJ4cLk27F6RSS44LAOidNFt1u4OfzhPXEqdNS84FTkxbFyfKCxOuRfMXJoc7s5ZkFuLkcOG5QHLMOvVcIPwWBOnMmBxO/0v1HieLxYKQkBAAQI8ePVBTUwMASEhIQHNzs+u7IyIiIvIgqpfqIiIiMG3aNMTGxqK8vBwDBgwAAGRnZyM8PNwtDRIRERF5CtWFU0FBAcrKylBTU4OkpCSkpqYCAEwmE8aMGeOWBomIiIg8heqlOoPBgKSkJFRVVcHPz8/28+joaMyapX0dnIiIiKg9EX3Jr8ViQXJysu1xYmIiAzCJiIjopqN6qa5NcHAwzGYz4uLiYLVaUVlZCZPJ5OreiIiIiDyKaOFkNptRWlqKiooKGI1G9O3bFykpKa7ujYiIiMijiJLDiYiIiEh4jxMRERERCS/VXa8ojbTs/VfSsrUSax1KqxWkNUv2B8jSdgFZKnJUgkba99bWtG9J/32G5Khuc+CzIgCyuWq9RkDr6ySemWQW0uNCsF1UgiDteKv8+HHGzBw5rsXHhTR5W3D8SGcR+ZD6F3of+qRQPAvxMZuoMY/NC5x6LEoT/Z05f3H/ktdS2JfoWNSYPXBl/hrvuf1bW/uXHNvi/p14LnDme47J4TcXfuJEREREJMSFExEREZGQ6sJpy5Yttj+fPXsWhYWFSE9PR2FhIc6cOePy5oiIiIg8ierC6bXXXrP9ubCwEEFBQcjPz0dYWBiys7Nd3hwRERGRJxHfHH769Gk8//zzAICwsDB8+OGHLmuKiIiIyBOpLpwaGxttl+s6duyIQ4cOITIyErW1tfjxxx/d0iARERGRp1BdOEVHR+Ojjz4CAHTr1g1nz54FACxevBgzZsxwfXdEREREHkSUHK4oChobG6EoCgICAtzRFxEREZHHUf3E6dixYzCbzThx4gRqa2sRFhaGpqYmREVFISsrC0FBQaKdODO0Mube6arb7KtaAgCIiXtOfbs9zyP6PvVPzb6uXNxa6x6NfX5xZZ/9NPa5+3lxUGBMf/Xe9u1ajKjBGgFuWxZcqaXR/64lmvMCrsxM+FpKZiGdq2Sf0QO0PwH9eudiUV+AbGbS40d0XAv2B0DzeX6988oxK3ifSN4jANBnqEZoZVmR+H0pmQXgnNe8bRbOrCV5XwLOPf4l+xS/l5wQ0gtcCa104mspnoXkuNaYF3BlZsL3nOS1lAZ4MgCzfVD9rbq8vDzk5ORg48aNWL9+PWJiYrBp0yaMGjUK06erH3RERERE7Y3qwslisSAkJAQA0KNHD9TU1AAAEhIS0Nzc7PruiIiIiDyI6qW6iIgITJs2DbGxsSgvL8eAAQMAANnZ2QgPD3dLg0RERESeQnXhVFBQgLKyMtTU1CApKQmpqakAAJPJhDFjxrilQSIiIiJPoXqpzmAwICkpCVVVVfDz87P9PDo6GrNmaX8DNREREVF7IvqSX4vFguTkZNvjxMREtLS0uKwpIiIiIk8k+sqV4OBgmM1mxMXFwWq1orKyEiaTydW9EREREXkUUQDmpUuXUFpaigMHDsBoNCI6OhopKSnw9vZ2R49EREREHkG0cCIiIiIi4aW66yVODpckzErTXgVJtOIUbOk+BdtFxavfVL+/fOGVWtppwX0ezFbd5sDn8wFAlHAtTdsVz8wJCb9tKcySVF6tFGOgNclYnDYt2E58XGukxX+9zezUFHJAmHwunEXvpNmq2x38dJ58roJZALLn6cxvJJCfC4THrOT4ceKxKE5uF8w/KkH7F3/2b10oTw4XvH/FcxW8T7ReS+BKWrwTX3PpcS2pxeRwzye6OZyIiIiIuHAiIiIiElNdOOXm5mLfvn3u6oWIiIjIo6ne41RdXY1Lly7hlVdeQVpaGvr37++uvoiIiIg8jurCyc/PD/Pnz8exY8fw+uuvo6ioCLGxsYiMjIS/vz9GjBjhrj6JiIiIbjjVhZPBYAAA9OzZE3l5eWhpaUFVVRX27duHY8eOceFERERENxXVhdOtt9561WNvb28MGjQIgwYNcmlTRERERJ5IFICpKAoaGxuhKAoCAgLc0RcRERGRx1H9xOnYsWMwm82oq6vDd999h7CwMDQ1NSEqKgpZWVkICgoS7UQaDqYVvOZQ6JpgO8n+AHmAZ5TG89y/zYyoxCz1bTYvEO+z9zCNYMJN81prSQIYhQGS4plJQgcFwXiAbK5agaeAY6GnkqBSaZhpVILGa75hB1D8AAAPWUlEQVR1gbwv6fwlr7lwn5EPzVHd7tAnhfLXUnj8S56nM9+/znyPA8KgRul5UbBP6XlRUqv30BzVbQDgYFmR+LWUBPWKg1EFx5nW+xJofW86dZ+C97i4lrB/BmDeOKpxBHl5ecjJycH777+P9evXIyYmBps2bcKoUaMwfbr2P1RERERE7YnqwslisSAkJAQA0KNHD9TU1AAAEhIS0Nzc7PruiIiIiDyI6qW6iIgITJs2DbGxsSgvL8eAAQMAANnZ2QgPD3dLg0RERESeQnXhVFBQgLKyMtTU1CApKQmpqakAAJPJhDFjxrilQSIiIiJPoXqpzmAwICkpCVVVVfDz87P9PDo6GrNmad/ARkRERNSeiL7k12KxIDk52fY4MTERLS0tLmuKiIiIyBOpXqprExwcDLPZjLi4OFitVlRWVsJkMrm6NyIiIiKPIgrAvHTpEkpLS3HgwAEYjUZER0cjJSUF3t7e7uiRiIiIyCOIFk5EREREJLxUd70kybGALFVYnFAsSh4Wpr06Mfk8arDGPrdcSU6WJPw6MTlca39t+5TOzJnzF6V4a8wVaJ2tNPlZ8jqJ+xcc1+L3iLR/wcykadm9kvNUt6v5oEDcl/SYlcxD+l4S1XLicQE4N/lcVEuaYu+EYxFonZn4mHXqLLTfc1rPEWh9ntKZSVLZnToL4blMuk8mhzuf6OZwIiIiIuLCiYiIiEhM9VKd1WrFhx9+iG3btqGhoQGKoqB79+548MEHMXjwYHf1SEREROQRVBdO+fn5CA4ORmpqKrZt2wZFUdC3b19s2LABlZWVyMzUvi+GiIiIqL1QXTgdO3YMc+fOBQDExsbiqaeewsSJE/HAAw9g9OjRbmmQiIiIyFOo3uOkKAq2bduGpqYmvPfee7jlllsAAFu2bHFLc0RERESeRPUTp7lz52LJkiU4fvw4evXqhYKCAgDAV199hUWLFrmlQSIiIiJPIQrAtFqtOHv2LBRFQUBAgDv6IiIiIvI4mvc4mc1mnDhxArW1tQgLC0NTUxOioqKQlZWFoKAgd/VJREREdMOp3uOUl5eHnJwcbNy4EevXr0dMTAw2bdqEUaNGYfr06e7qkYiIiMgjqC6cLBYLQkJCAAA9evRATU0NACAhIQHNzc2u746IiIjIg6heqouIiMC0adMQGxuL8vJyDBgwAACQnZ2N8PBwtzRIRERE5ClUbw5XFAVlZWWoqamBv78/UlNTAQDFxcUYM2YMAgMD3dYoERER0Y2meqnOYDAgKSkJVVVV8PPzs/08Ojoas2apf8szERERUXsj+pJfi8WC5ORk2+PExES0tLS4rCkiIiIiT6R6j1Ob4OBgmM1mxMXFwWq1orKyEiaTyeGdNTc345FHHsEzzzyDUaNG4ciRI8jNzYXBYECPHj2Qn58PLy/1lnbu3IkpU6bgrrvuAtB6H9acOXMAAK+//jrMZjN27doFHx8fUU/26n355ZdYtGgRvLy80LFjRyxevBj+/v66as2aNQv79+/Hb3/7WwDAuHHjkJiYqKvW5MmT0djYCAA4e/Ys7r77bhQWFuqqpWf+ALBx40a8+uqr8PLywuTJk23PRc/87dXTM397tfTM314tPfO3V0vP/N99911s3LjR9vjrr7/Gl19+CavViqVLl2LdunWorKzU7EerXllZGVatWgVvb2/4+/tj8eLF6NSpk65a6enpuHjxIrp06QIAyMzMRHR0tMO1vvjiCzz11FO2n508eRKPP/44JkyYoKuvPXv2YOHChfD29ka/fv0wbdo01ToA8MMPPyAzMxNNTU1oaWnBxIkTER8fr2v+9mrpmb1aPT3zv1atQYMG6Zq/vb70zN9qtSIvLw/ffvstvL29kZ+fj7CwMACOn3/s1dJ77rFXT8/5x14tPecfe7X0nv8JgCLQ0tKivPPOO0p+fr5SWFiolJaWKhaLRfKfXmXp0qXKqFGjlPXr1yuKoigTJkxQNm/erCiKohQXFysbN27UrFFZWak8++yzv/h5aWmpsnTpUiUxMVG5cOGCuCd79Z599lnlP//5j6IoirJixQpl5cqVumtlZmYqn332mbgntVo/N2vWLGXv3r26a+mZ/5kzZ5SHHnpIOX/+vFJfX6/Mnj1bURT987dXT8/87dXSM397tX5OOn97tfTM/+d27typ5OfnK4qiKCtXrlTeeOMNpX///g7VsFcvIyNDOXfunKIorc/zenpLS0tTampqnNLXz40bN045ceKE7lqPPvqo8u233yqKoihZWVnK7t27Nf/7kpISZcmSJYqiKMr333+vPPzww4qi6Ju/vVp6Z2+vnp7526v1c9L526ulZ/6ffPKJMmXKFEVRFOX48ePK+PHjFUXRd/6xV0vPuUetnp7zj71aPyc9/9irdb3nn5uZaHnp5eXl0Jf6trS0YNasWairq0OnTp2waNEiXLhwAYcPH75qpX38+HHExsYCAOLj47F27Vo8+uijqrWeeOKJa+4zKSkJvr6+eP/99x3qzV695cuXA2i9Qb6+vh79+vXTXUvC0VpHjx7F+fPnbfPTU0vP/P/whz9g4MCB8PX1ha+vr+3/dvTO3149PfO3V0vC0VqOzN9eLT3zX7RokS149sUXX8SSJUsAAGlpafD19bXNTfo87dVbs2YNAODSpUs4derUNcNupbUkHK1VUVGBHj16IDg4WHetU6dO2X47+IEHHsD27dsRFxenWuuJJ56wxbKcO3cOt956KwDZ/KW1JLN3pJ6Eo7Ucmb+9Wnrmf++999reM6GhoThx4gQuX74sOv9Ia0nOPY7Uk5DWMhqNABw7/9irJTn/0LWJ7nFy1D//+U9069YNb731FsaMGYOysjKYzeZf3FAeERFh+8Lg8vJynD59WrNW2wJswoQJSE1Nxfbt2wEAvr6+unqzVw8Atm7diuHDh+P06dMYOXLkddV64403kJGRgalTp+LMmTPXVQto/Vg6LS3tup6jnvnX1dWhubkZEyZMwJNPPokdO3YA0D9/e/UAx+evVsvR+avVAhybv71aeuZfVlYGoPX7IoODg3HbbbcB0D9/e/UAYMOGDUhKSkJoaCj69+9/XbWWL1+OsWPHIjc395oZcI7UAlrnn5GRcV3P8Y477kBVVRUURUFFRYX4/HPixAkMGzYMaWlpyMzMBCCbv7QWoD17R+s5On+1WoBj87dXS8/8/f39sW3bNly+fBlHjx5FbW0tGhsbdc3fXi1A+9zjaD1Hzz9qtdrmLz3/2KslOf/Qtblk4bR//37b/zmkpKSgS5cuuPvuu21hmm0yMzPx4YcfIiMjA4qiQLlGMsL/1ho6dCgmTZqElStXwmw2IycnBxaLRXdvavUSEhLw0Ucf4c4778SqVat013rssccwffp0vP766+jduzeKi4uvqy+LxYLdu3fjvvvuu67nqGf+HTt2xNmzZ1FcXIyFCxciKyvrmv+ddP5q9Rydv71aeuav1pej87dXS8/8n3zySQDAunXr8Pjjj4vnrqfeqFGj8Omnn6Kpqema/ycvrZWRkYGZM2fizTffhMFgwJtvvnldfdXX1+PixYsIDQ29rudYVFSE4uJijBs3Dl27dhXN38fHByaTCZs2bcKaNWswd+7ca/Yg6UutltbsHamnZ/5qvTk6f3u19Mz/ySefRExMDMaOHYs1a9bgzjvvFJ9/HKmlde5xpJ6e849ab46ef+zVkpx/6NpccieY0WiE1Wq1Pd68eTNqa2uxefNmfP/99+jYsSNuv/12DBo0CC+//DKA1hXvyZMnNWsFBQXZfsMvNDQU3bp1Q319/S8WZdLe7NU7dOgQhg0bBoPBgIcffhgrVqzQXWvgwIG2bYYMGYL8/Pzrep5VVVXX/IhWTy1H5x8QEIDf//738PLyQmhoKHx8fHDmzBnxlz9L6+3Zs8fh+durpWf+as/T0fnbqxUcHOzw/Nvs3LkTs2fPttuDtLdr1fvpp5+wc+dOJCQkwMvLC0OHDsWuXbt+8TG+tLdhw4bZ/jxkyBB88MEHuvpqs2XLFrv/aDhSKyIiwnZZ7K233sK5c+c0a+3ZswcPPPAAACAyMhInT5686hKKGkmtixcv4osvvtCcvSO96Zm/2vN0dP72aumZPwBMnTrV9uekpCTd5x57tTZt2qR57nGk3s8/LZWef+zVAuDw+cderQ4dOmief+jaXPKJU0xMjO03Sz7//HNERkZi/fr1eOeddzB69Gg888wzGDRoEJYvX47NmzcDaP1oesiQIZq1Vq5ciddeew1A6zXyhoYGh75sWFpvxYoVOHjwIABg79696Nmzp+5azz77LGprawG0nrzbfrtN7/Pct28fIiMjr/s56pn/yZMnUVlZCavVisbGRly8eNGheymk9fTM314tPfNXe56Ozt9eLT3zf+mll1BfXw8fHx907NhRe+A66hmNRsyZMwf19fUAWi9xSeZ/rVqKouCpp56y/aMonb/a83R0/vZqZWVl4dChQ7h8+TLee++9a/6m0//Wuv3227F3714AQF1dHXx8fESLJmmtjh07imYvrdehQwdd81d7no7O314tPfMvLi5GVlYWgNbLaX369EGHDrJ/xqS1JOceR+rpOf+oPU9H52+vluT8Q9emmhyul8ViwezZs3HixAl4eXnBbDbb/tFfsWIFunfvjlGjRuHo0aOYOXMmFEXBPffcY3tx1WoVFRWhsLAQ586dQ0tLCyZNmoTBgwdj5cqVqKioQHV1NWJiYnD33Xdj5syZuuvt27cPRUVFMBqNuOWWW7Bo0aJf/J+NtFZlZSUWL16Mzp07o0uXLliwYIHuWgBQWFiIfv36XZWtpaeWnvmbzWZ8/vnnWLduHQDg6aefxtChQ3XP3149PfO3V0vP/O3V0jN/e7X0zv/UqVP429/+hldffdW2XWFhIb755hvs2bMHcXFxGDJkCP785z/rrrdlyxasWLECHTt2RLdu3WA2m9G5c2ddtT744AO8+uqr6Ny5M4KCglBUVKS7FgBMmDABTz/9NPr27Sue/7Vq7d69G/PmzQMAPPLIIxg3bpxmraKiIixatAgNDQ24dOkSpkyZgoEDB4rmL60lmb0j9fTM314tPfO3V0vP/M1mM5YtW4bDhw+jU6dOWLJkCYKDg0XnH2ktybnHkXp6zz/XqgXoO/9cq5bk/EPX5pKFExEREVF75JJLdURERETtERdOREREREJcOBEREREJceFEREREJMSFExEREZEQF05EREREQlw4EREREQlx4UREREQk9P8AB3oK9liFet0AAAAASUVORK5CYII=\n",
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "tags": []
+ }
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "2kzVUPeF2NrG"
+ },
+ "source": [
+ "##Base Resultados"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "DsCTrRSI2UQD"
+ },
+ "source": [
+ "Como já colocamos no `paredão` a ***base de experimento*** e exploramos, vamos colocar no `paredão` a ***base de resultados*** para explorar."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "dPb5EVv-3auK"
+ },
+ "source": [
+ "Como fizemos na exploração anterior vamos ver se essa base tem a mesma **dimensão** que a base de experimento, para isso vamos utilizar o `.shape`"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "pORhjwKK2SuK",
+ "outputId": "bebedbac-893f-476c-9574-faa81e0f1478"
+ },
+ "source": [
+ "df_resultados.shape"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "(23814, 207)"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 162
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "8oh2U1jB5lBc"
+ },
+ "source": [
+ "Diferente da base de dados anterior, essa base tem menos colunas, um total de `207 colunas`. Vamos verificar as primeira linhas."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 261
+ },
+ "id": "ucHK3IoW2C5b",
+ "outputId": "f37f19aa-ea0f-4898-bb99-fc94d5b2be11"
+ },
+ "source": [
+ "df_resultados.head()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ ""
+ ],
+ "text/plain": [
+ " id ... wnt_inhibitor\n",
+ "0 id_000644bb2 ... 0\n",
+ "1 id_000779bfc ... 0\n",
+ "2 id_000a6266a ... 0\n",
+ "3 id_0015fd391 ... 0\n",
+ "4 id_001626bd3 ... 0\n",
+ "\n",
+ "[5 rows x 207 columns]"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 163
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "nFJFWfJu6NYh"
+ },
+ "source": [
+ "Vamos olhar os nomes das colunas e ver se conseguimos tirar alguma informação"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "05T4ix7742IV",
+ "outputId": "be7a4157-fc54-4dff-ed11-3a51136f47c8"
+ },
+ "source": [
+ "df_resultados.columns"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "Index(['id', '5-alpha_reductase_inhibitor', '11-beta-hsd1_inhibitor',\n",
+ " 'acat_inhibitor', 'acetylcholine_receptor_agonist',\n",
+ " 'acetylcholine_receptor_antagonist', 'acetylcholinesterase_inhibitor',\n",
+ " 'adenosine_receptor_agonist', 'adenosine_receptor_antagonist',\n",
+ " 'adenylyl_cyclase_activator',\n",
+ " ...\n",
+ " 'tropomyosin_receptor_kinase_inhibitor', 'trpv_agonist',\n",
+ " 'trpv_antagonist', 'tubulin_inhibitor', 'tyrosine_kinase_inhibitor',\n",
+ " 'ubiquitin_specific_protease_inhibitor', 'vegfr_inhibitor', 'vitamin_b',\n",
+ " 'vitamin_d_receptor_agonist', 'wnt_inhibitor'],\n",
+ " dtype='object', length=207)"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 164
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "fvE2mHd744JV",
+ "outputId": "89d84d8b-1c08-44bd-e43f-af22e0f694ae"
+ },
+ "source": [
+ "#Verificar os tipos de dados das colunas do dataset\n",
+ "df_resultados.info()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "![\"imersão](\"https://media1.tenor.com/images/9a883dd19f6ac5c7cb1ae882d458afcd/tenor.gif?itemid=21270823\")
\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "EpNZeXFy-hAW"
+ },
+ "source": [
+ "Mas calma, ainda temos muito trabalho pela frente, a vida de um Cientista é assim mesmo, precisamos testar, testar e no fim testar!"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "1iU1lVbr-i_X"
+ },
+ "source": [
+ "Vamos criar uma matrix para entender como está o erro, a coluna `Real` é a coluna dos **dados certos**, ja a coluna `Predictions `é a coluna que a **máquina previu**."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "XDLhIdgF-uot",
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 402
+ },
+ "outputId": "588f27b5-35f9-4ab1-ad11-0e5edcdfdfba"
+ },
+ "source": [
+ "mtx = pd.DataFrame({'Real':y_test,'Predictions': y_predict})\n",
+ "mtx"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/html": [
+ "\n",
+ "\n",
+ ""
+ ],
+ "text/plain": [
+ " Real Predictions\n",
+ "8579 1 0\n",
+ "16718 0 0\n",
+ "17007 1 1\n",
+ "1901 1 1\n",
+ "13470 1 1\n",
+ "... ... ...\n",
+ "619 0 1\n",
+ "12019 1 0\n",
+ "16869 1 1\n",
+ "7536 1 0\n",
+ "1551 0 1\n",
+ "\n",
+ "[7145 rows x 2 columns]"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 214
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "JRgqwrE--zOW"
+ },
+ "source": [
+ "Vamos importar a biblioteca que mostra várias métricas de Classificação. Para isso utilizamos a biblioteca `sklearn.metrics` e importamos o `classification_report`"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "d2GN4X5r-w9o",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "c8ba3074-26cd-4e4c-e85e-6233cb9f810c"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.54 0.40 0.46 2841\n",
+ " 1 0.66 0.77 0.71 4304\n",
+ "\n",
+ " accuracy 0.62 7145\n",
+ " macro avg 0.60 0.59 0.59 7145\n",
+ "weighted avg 0.61 0.62 0.61 7145\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "zOPgJvxl_Yc-"
+ },
+ "source": [
+ "Assim ele mostra todas as métricas importantes para avaliar o modelo. \n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "6FwF86gIB4HS"
+ },
+ "source": [
+ "62% significa que a cada 100 previsões ele irá acertar apenas 62 e errar 38 de todas as 100, sendo muito baixo."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "4K4zvJ4lCMzF"
+ },
+ "source": [
+ "Esse modelo não está muito bom, vamos ver outro modelo e eu escolho o XGB."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "RiN9MEjH_UTE"
+ },
+ "source": [
+ "##Modelo XGBoost antes do Balanceamento"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "cdIYsTt7AH9V"
+ },
+ "source": [
+ "Agora vamos utilizar o XGB, para isso precisamos importar a biblioteca `xgboost` e escolher o modelo de `XGBClassifier`.\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "7XNWlATD_79V"
+ },
+ "source": [
+ "\n",
+ "\n",
+ "Para ver a documentação [Clique Aqui](https://xgboost.readthedocs.io/en/latest/python/python_api.html)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "MLWBSjZtAj-l"
+ },
+ "source": [
+ "Vamos instanciar o modelo e treinar com os dados"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "zKVX5NZ3_gxz"
+ },
+ "source": [
+ "model_x=XGBClassifier()\n",
+ "model_x.fit(x_train,y_train)\n",
+ "y_predict=model_x.predict(x_test)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "dQTuAQjz_t8D"
+ },
+ "source": [
+ "###Avaliando o Modelo"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "4hqnLMg0ApQW"
+ },
+ "source": [
+ "Agora que treinamos vamos **avaliar** como fizemos com o modelo anterior e irei fazer a `avaliação` nos `demais modelos`, para poder `comparar` depois."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "o3BO6x1k_mYT",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "8552f4af-f1cd-487f-8be7-b742d8a20a9b"
+ },
+ "source": [
+ "model_x.score(x_test, y_test)*100"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "65.47235829251224"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 217
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "Eviu6JWa_jBz",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "52facb3b-2aa6-4fcf-a3ff-b8ed5f482123"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.67 0.26 0.38 2841\n",
+ " 1 0.65 0.91 0.76 4304\n",
+ "\n",
+ " accuracy 0.65 7145\n",
+ " macro avg 0.66 0.59 0.57 7145\n",
+ "weighted avg 0.66 0.65 0.61 7145\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "7BeAq3TVBbK_"
+ },
+ "source": [
+ "Esse modelo está com um Score de 65%, aumentou 3% referente ao anterior, melhor que o Regressão Logistica, mas ainda não é o suficiente. \n",
+ "\n",
+ "Então vamos tentar mais um modelo, caso não seja bom, vamos tentar descobrir como podemos melhorar."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "x4coXQE__3F4"
+ },
+ "source": [
+ "##Modelo RandomForest antes do Balanceamento"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "IW3B5aMVCaSK"
+ },
+ "source": [
+ "Agora usaremos o Random Forest, para utilizar ele precisamos importar a biblioteca `sklearn.ensemble` e escolher o modelo de `RandomForestClassifier`.\n",
+ "\n",
+ "Para acessar a documentação [clique Aqui](https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.RandomForestClassifier.html)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "M5UrVmgV___L"
+ },
+ "source": [
+ "#instaciando o modelo e treinando\n",
+ "model_rf = RandomForestClassifier(n_estimators=100,random_state=SEED, n_jobs=-1)\n",
+ "model_rf.fit(x_train, y_train)\n",
+ "y_predict = model_rf.predict(x_test)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "eUDyo19DAKCy"
+ },
+ "source": [
+ "###Avaliando o Modelo"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "dZu0ujmIDcme",
+ "outputId": "72701a3b-4d0e-4910-ab1c-e6b2f3c74b22"
+ },
+ "source": [
+ "model_rf.score(x_test, y_test)*100"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "64.35269419174247"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 225
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "FhyFkSvcAIPi",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "0f21c5a0-6855-4aa1-d431-2d43fbf64e11"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.62 0.26 0.37 2841\n",
+ " 1 0.65 0.90 0.75 4304\n",
+ "\n",
+ " accuracy 0.64 7145\n",
+ " macro avg 0.64 0.58 0.56 7145\n",
+ "weighted avg 0.64 0.64 0.60 7145\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Um4es6S6Dl9e"
+ },
+ "source": [
+ "O Random Forest ainda ficou pior que o XGBoost, com 64%. \n",
+ "\n",
+ "Na nossa `analise de exploratória`, vimos que a coluna que `com_droga` está `muito desbalanceada`, vamos balancear ela e verificar se `melhoramos` o score dos `nossos modelos`"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "_2_uO7c1AXmH"
+ },
+ "source": [
+ "#***Balanceando os Dados***"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "5Slq9I8_E6Q9"
+ },
+ "source": [
+ "Desbalanceamento ocorre quando temos um conjunto de dados que possui `muitos exemplos` de uma classe e `poucos exemplos` da outra classe. Como vimos na coluna `com_droga`.\n",
+ "\n",
+ "Quando isso ocorre, temos algumas técnicas para aplicar , por exemplo o Upsample e Downsample. "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "RIUKkolyEfg3"
+ },
+ "source": [
+ "***Upsample:***\n",
+ "\n",
+ "\"Upsampling é um técnica de processamento digital de sinais para aumentar artificialmente a taxa de amostragem em N vezes, inserindo um número N-1 de zeros entre as amostras originais do sinal, e passando o conjunto obtido por um filtro de reconstrução, que nada mais é que um filtro do tipo passa-baixas. Para mostrar essa técnica utilizaremos o programa anterior ligeiramente modificado. O script a seguir simula o upsampling da senoide ideal e realiza a filtragem de reconstrução.\"\n",
+ "\n",
+ "***Downsample:***\n",
+ "\n",
+ "\"Downsampling ou decimação é a técnica de redução da taxa de amostragem. Isso é feito simplesmente separando uma amostra a cada N. Há uma inevitável distorção do sinal.\"\n",
+ "\n",
+ "Fonte: [Clique Aqui](https://www.embarcados.com.br/oversampling-upsampling-downsampling-dsp/#:~:text=A%20sobre%2Damostragem%20(oversampling),artificial%20da%20taxa%20de%20amostragem.)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "t35_Sx_xEDku"
+ },
+ "source": [
+ "![\"imersão](\"https://www.mdpi.com/education/education-09-00275/article_deploy/html/images/education-09-00275-g001-550.jpg\")
\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "JbMGkArYGMdd"
+ },
+ "source": [
+ "Vamos verificar nossa coluna e a frequencia de seus valores"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "XFZO4asAA0aM",
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 380
+ },
+ "outputId": "03968605-1b6f-47ff-aa36-76248716f0b3"
+ },
+ "source": [
+ "sns.countplot(x='com_droga', data = df, palette=\"crest\")\n",
+ "plt.title('Tipos de Tratamentos', fontdict={'fontsize':18, 'fontweight': 'bold'},)\n",
+ "plt.xlabel('Tratamento') \n",
+ "plt.ylabel('Quantidade') \n",
+ "plt.show()"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "display_data",
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ "![\"tiago\"](\"https://media1.tenor.com/images/aeb09d020883abe4713264ac4a4ce748/tenor.gif?itemid=16826989\")
\n",
+ "\n",
+ "\n",
+ "Antes sem o balanceamento da coluna `com_droga` o modelo de `Regressão Logística` tinha um score de 62% e só com o balanceamento subiu para 78%, isso é **muito bom**... estamos no **caminho certo**!\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "kAdQ-iWnCDEs",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "eb6acf94-6f9b-4fcb-9179-08f3beee3972"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.82 0.86 0.84 8776\n",
+ " 1 0.69 0.62 0.65 4393\n",
+ "\n",
+ " accuracy 0.78 13169\n",
+ " macro avg 0.76 0.74 0.75 13169\n",
+ "weighted avg 0.78 0.78 0.78 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "j95dH3TXLm81"
+ },
+ "source": [
+ "Agora vamos testar o modelo que se saiu melhor com a coluna desbalanceada o XGBoost"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "ONT8QS25CqIb"
+ },
+ "source": [
+ "##Modelo XGB depois do Balanceamento"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "vXASElvfCJBT"
+ },
+ "source": [
+ "#Instanciando o modelo e treinando\n",
+ "model_x=XGBClassifier()\n",
+ "model_x.fit(x_train,y_train)\n",
+ "y_predict=model_x.predict(x_test)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "navZD2kgCdli"
+ },
+ "source": [
+ "###Avaliando o Modelo"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "RCn1-jWoCOVN",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "1010b770-5b3b-4ad8-ee5b-c47394b56a24"
+ },
+ "source": [
+ "model_x.score(x_test, y_test)*100"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "77.86468220821627"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 248
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "aFyrvuEgCXoy",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "071611e6-a632-4c4e-c7bc-c7543f4564bd"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.82 0.85 0.84 8776\n",
+ " 1 0.68 0.63 0.65 4393\n",
+ "\n",
+ " accuracy 0.78 13169\n",
+ " macro avg 0.75 0.74 0.75 13169\n",
+ "weighted avg 0.77 0.78 0.78 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Hqx7eES9MvEo"
+ },
+ "source": [
+ "Esse também melhorou depois do balanceamento, antes era `65%` e agora está com `77%`, teve uma melhora de 123%, porém o modelo de Regressão Linear está melhor que ele. "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "lZ88Ga6bNPXv"
+ },
+ "source": [
+ "O que mais podemos fazer para melhorar nossas métricas?\n",
+ "\n",
+ "Vamos utilizar a técnica de Feature Seletion."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Y48QXabhD-Pg"
+ },
+ "source": [
+ "#***Feature Selection (Seleção de Variáveis)***"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "_WTkfjPPETBQ"
+ },
+ "source": [
+ "Há boatos que para criar um modelo de aprendizado de máquina precisa de muita calma, paciencia e muito teste. Realmente, você precisa saber diversas técnicas, como elaborar, construir e refinar o seu modelo e testar! \n",
+ "\n",
+ "Uma dessas técnicas é a Seleção de Variáveis. Assim como o nome já diz, essa técnica seleciona as melhores variáveis para o modelo, ter variáveis de mais isso pode prejudicar a performance do algoritmo. \n",
+ "\n",
+ "Muitas features podem causar problemas como duração para treinamento do modelo ou dificuldades de colocar modelo de produção. Essa técnica ajuda a reduzir o overfitting e aumenta a accurácia do modelo e como já foi falado reduz o tempo de treino. \n",
+ "\n",
+ "Se quiser saber mais sobre essas técnicas em python [Clique Aqui](https://medium.com/data-hackers/como-selecionar-as-melhores-features-para-seu-modelo-de-machine-learning-faf74e357913)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "9sG8T_tzPKGN"
+ },
+ "source": [
+ "Para isso vamos utilizar o *SelectFromModel*, importanto a biblioteca `sklearn.feature_selection` e selecionar o `SelectFromModel`\n",
+ "\n",
+ "para acessar a documentação [clique aqui](https://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.SelectFromModel.html)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "a56qrp26Pv0p"
+ },
+ "source": [
+ "Vamos utilizar o modelo RandomForest para classificar as melhores variáveis, depois vamos treinar com RandomForest também para verificar as métricas."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "IQWvhPwpQMlV"
+ },
+ "source": [
+ "Para isso utilizei uma estrutura de repetição chamada For, para verificar em 10 e 10 quais as melhores variáveis.\n",
+ "\n",
+ "Deixei o código comentado pois ele demora muito para executar!"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "lkcLBgbwQbOu"
+ },
+ "source": [
+ "\n",
+ " '''k_vs_score = []\n",
+ "#for k in range(10, x_train.shape[1], 10):\n",
+ " model_rf = RandomForestClassifier(n_estimators=100,random_state=SEED, n_jobs=-1)\n",
+ " selector = SelectFromModel(model_rf, max_features=40, threshold=-np.inf)\n",
+ " selector.fit(x_train, y_train)\n",
+ "\n",
+ " x_train2 = selector.transform(x_train)\n",
+ " x_test2 = selector.transform(x_test)\n",
+ "\n",
+ " mdl = RandomForestClassifier(n_estimators=100,random_state=SEED, n_jobs=-1)\n",
+ " mdl.fit(x_train2, y_train) \n",
+ " p = mdl.predict(x_test2)\n",
+ "\n",
+ " score = mdl.score(x_test2, y_test)\n",
+ " print(\"k = {} - MAE = {}\".format(k, score))\n",
+ "\n",
+ "\n",
+ " mask = selector.get_support()\n",
+ " print(Xtrain.columns[mask])\n",
+ " k_vs_score.append(score)\n",
+ "'''"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "gKvgdmTgQpB2"
+ },
+ "source": [
+ "Após executar todo o código, foi observado que o melhor `valor` e onde o `score` está mais estável é quando o` K` é igual a `40`, `k` significa o número de variáveis,conforme imagem abaixo. "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "tt-B3hSdSZbd"
+ },
+ "source": [
+ ""
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "y62TjDIMSj__"
+ },
+ "source": [
+ "Vamos utilizar o `get_support() `para criar uma mascara e pegar as melhores variáveis que o `SelectFromModel` selecionou."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "CYy7iEETEMXC",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "da21f415-c36f-47f6-9444-35d1c01e1197"
+ },
+ "source": [
+ "model_rf = RandomForestClassifier(n_estimators=100,random_state=SEED, n_jobs=-1)\n",
+ "selector = SelectFromModel(model_rf, max_features=40, threshold=-np.inf)\n",
+ "selector.fit(x_train, y_train)\n",
+ "\n",
+ "x_train2 = selector.transform(x_train)\n",
+ "x_test2 = selector.transform(x_test)\n",
+ "\n",
+ "mdl = RandomForestClassifier(n_estimators=100,random_state=SEED, n_jobs=-1)\n",
+ "mdl.fit(x_train2, y_train) \n",
+ "p = mdl.predict(x_test2)\n",
+ "\n",
+ "score = mdl.score(x_test2, y_test)\n",
+ "print(\"k = {} - Score = {}\".format(40, score))\n",
+ "\n",
+ "\n",
+ "mask = selector.get_support()\n",
+ "print(x_train.columns[mask])"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "k = 40 - Score = 0.8124383020730503\n",
+ "Index(['g39', 'g50', 'g51', 'g57', 'g58', 'g68', 'g75', 'g100', 'g122', 'g138',\n",
+ " 'g175', 'g178', 'g206', 'g223', 'g230', 'g312', 'g317', 'g322', 'g328',\n",
+ " 'g352', 'g365', 'g367', 'g418', 'g445', 'g463', 'g524', 'g525', 'g529',\n",
+ " 'g600', 'g620', 'g635', 'g656', 'g689', 'g701', 'g738', 'g764', 'c13',\n",
+ " 'c65', 'c73', 'c98'],\n",
+ " dtype='object')\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "-PzoF46ZTJR2"
+ },
+ "source": [
+ "Esses são os nomes das colunas que o `SelectFromModel ` selecionou para ser as *melhores* `variáveis preditoras` do modelo."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "lQPIWjRWE2X_"
+ },
+ "source": [
+ "[['g39', 'g50', 'g51', 'g57', 'g58', 'g68', 'g75', 'g100', 'g122', 'g138',\n",
+ " 'g175', 'g178', 'g206', 'g223', 'g230', 'g312', 'g317', 'g322', 'g328',\n",
+ " 'g352', 'g365', 'g367', 'g418', 'g445', 'g463', 'g524', 'g525', 'g529',\n",
+ " 'g600', 'g620', 'g635', 'g656', 'g689', 'g701', 'g738', 'g764', 'c13',\n",
+ " 'c65', 'c73', 'c98']]"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "fi_SaOdCTism"
+ },
+ "source": [
+ "Agora vamos colocar no x as nossas novas variáveis preditoras"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "g_dscgJ_EVtH"
+ },
+ "source": [
+ "x = dados[['g39', 'g50', 'g51', 'g57', 'g58', 'g68', 'g75', 'g100', 'g122', 'g138',\n",
+ " 'g175', 'g178', 'g206', 'g223', 'g230', 'g312', 'g317', 'g322', 'g328',\n",
+ " 'g352', 'g365', 'g367', 'g418', 'g445', 'g463', 'g524', 'g525', 'g529',\n",
+ " 'g600', 'g620', 'g635', 'g656', 'g689', 'g701', 'g738', 'g764', 'c13',\n",
+ " 'c65', 'c73', 'c98']]\n",
+ "\n",
+ "y = dados['atv_moa']"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "3OGyjUrQGTIi"
+ },
+ "source": [
+ "#**Criação de Maquina Preditiva após Balanceamento e Seleção de Variáveis**"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "RYlokH8cTy21"
+ },
+ "source": [
+ "Agora que já fizemos o balanceamento, a selecão de variáveis, vamos verificar se o modelo melhorou sua performance?\n",
+ "\n",
+ "![\"win\"](\"https://media0.giphy.com/media/EDt1m8p5hqXG8/giphy.gif?cid=ecf05e47j94osktw8orptmz3cwhhd7ohce977pomhn729nbu&rid=giphy.gif&ct=g\")
\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "APDeT3xhUqI1"
+ },
+ "source": [
+ "Agora vamos utilizar os mesmos modelos preditivos de Classificação: `Regressão Logistica`, `XGBoost` e `RandomForest`"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "_1GGvIrBF0UO"
+ },
+ "source": [
+ "##Regressão Logistica após Balanceamento e Selecão de Variáveis"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "f1fflXLMF4Jd"
+ },
+ "source": [
+ "#treinando o modelo\n",
+ "model_rl.fit(x_train, y_train)\n",
+ "y_predict = model_rl.predict(x_test)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "eQBfJW8uGBkS",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "8eb646ae-35dc-4fa7-ec1c-487116bdaa3e"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.82 0.86 0.84 8776\n",
+ " 1 0.69 0.62 0.65 4393\n",
+ "\n",
+ " accuracy 0.78 13169\n",
+ " macro avg 0.76 0.74 0.75 13169\n",
+ "weighted avg 0.78 0.78 0.78 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "mIK4HTTsFtxk"
+ },
+ "source": [
+ "##XGB após Balanceamento e Selecão de Variáveis"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "U5Xq193XFeu5",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "a13a69d1-6970-4a7a-c6b3-64e11ff0f8af"
+ },
+ "source": [
+ "model_x.fit(x_train,y_train)\n",
+ "model_x.predict(x_test)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "array([0, 0, 0, ..., 0, 0, 0])"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 254
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "5Tas6A6BFc__",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "84171fc4-1be7-4302-bbb9-34f993580e4d"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.82 0.86 0.84 8776\n",
+ " 1 0.69 0.62 0.65 4393\n",
+ "\n",
+ " accuracy 0.78 13169\n",
+ " macro avg 0.76 0.74 0.75 13169\n",
+ "weighted avg 0.78 0.78 0.78 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "rb2KC-rhFJ-j"
+ },
+ "source": [
+ "##Random Forest após Balanceamento e Selecão de Variáveis"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "qhsVButCEwbY"
+ },
+ "source": [
+ "mdl.fit(x_train, y_train)\n",
+ "y_predict = mdl.predict(x_test)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "grzOk3oFFCbq",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "cc66349d-32e4-4601-fbc1-a85fff98b181"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.89 0.81 0.85 8776\n",
+ " 1 0.68 0.80 0.73 4393\n",
+ "\n",
+ " accuracy 0.81 13169\n",
+ " macro avg 0.78 0.80 0.79 13169\n",
+ "weighted avg 0.82 0.81 0.81 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "sFfV8U1sW_Kt"
+ },
+ "source": [
+ "Vamos fazer uma comparação entre os modelos"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "T0ajI-ShVveV"
+ },
+ "source": [
+ "Modelo |Sem Balanceamento | Com Balanceamento | Com Seleção de Variáveis e Balanceamento |\n",
+ "-------------------|------------------|------------------|------------------\n",
+ "Regressão Logística | 62% | 78% | 78%\n",
+ "XGBoost | 65% | 77% | 78%\n",
+ "RandomForest | 64%| Não utilizado | 81%\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "VCvQ7SBCXB8E"
+ },
+ "source": [
+ "O modelo de RandomForest teve uma melhor métrica 81% de score.\n",
+ "\n",
+ "Portanto, o RandomForest ganhou o Objetivo 1!\n",
+ "![\"win\"](\"https://media1.tenor.com/images/0dc98cc3b67969a13cb04e1a110ad6f5/tenor.gif?itemid=17070545\")
\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "S0Ao3kotHxmr"
+ },
+ "source": [
+ "#***Criação da Maquina Preditiva***\n",
+ "Objetivo 2: Prever se o experimento foi tratado com Droga ou Com controle."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "4Gt65VCKXKL1"
+ },
+ "source": [
+ "Agora vamos partir para nosso segundo objetivo! \n",
+ "\n",
+ "Prever se o experimento foi utilizado droga ou controle."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "-HQLpNOVXSRd"
+ },
+ "source": [
+ "Para isso vamos selecionar nosso` X e Y`\n",
+ "\n",
+ "Nosso `y `será a coluna `com_droga` e vamos selecionar todas as variáveis que são inteiro ou float para ser nosso `x`\n",
+ "\n",
+ "após isso vamos fazer a divisão da base de dados em teste e treino, deixando 30% para testar."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "-2HVsErfID0T"
+ },
+ "source": [
+ "X = dados.select_dtypes('float64','int64')\n",
+ "y = dados['com_droga']\n",
+ "\n",
+ "X_train, X_test, y_train, y_test = train_test_split(x, y, test_size = 0.3, random_state = SEED)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "6iB40sXrXw_9"
+ },
+ "source": [
+ "Vamos utilizar nossos famosos modelos que já foram para o paredão diversas vezes no último objetivo. Quem será que vai ganhar esse objetivo? Pra quem vai a sua torcida?"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "sDCNDLRlIRn7"
+ },
+ "source": [
+ "##Regressão Logistica"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "5027OuDcIWeJ"
+ },
+ "source": [
+ "model_rl.fit(X_train, y_train)\n",
+ "y_predict = model_rl.predict(X_test)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "rfcpIlh2I-ZV",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "0f2ca93c-c938-4d89-fb64-2cbf8ea61014"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.85 0.88 0.87 6539\n",
+ " 1 0.88 0.85 0.86 6630\n",
+ "\n",
+ " accuracy 0.86 13169\n",
+ " macro avg 0.87 0.86 0.86 13169\n",
+ "weighted avg 0.87 0.86 0.86 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "Z3pdlB7WIqkB"
+ },
+ "source": [
+ "##XGBoost"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "isSL5qBNIp_C",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "16edc44a-eb67-4d89-9b9d-fef9e694119d"
+ },
+ "source": [
+ "#Treinando o modelo XGB\n",
+ "model_x.fit(X_train,y_train)\n",
+ "model_x.predict(X_test)"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "array([1, 0, 1, ..., 0, 0, 0])"
+ ]
+ },
+ "metadata": {
+ "tags": []
+ },
+ "execution_count": 263
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "W6I0bkDrJFKb",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "1ba1f091-1726-49e3-f473-62914f995219"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.85 0.88 0.87 6539\n",
+ " 1 0.88 0.85 0.86 6630\n",
+ "\n",
+ " accuracy 0.86 13169\n",
+ " macro avg 0.87 0.86 0.86 13169\n",
+ "weighted avg 0.87 0.86 0.86 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "8CwYWFE7Ig_K"
+ },
+ "source": [
+ "## RandomForest"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "d3uxPy0_Icqb"
+ },
+ "source": [
+ "#Treinando o modelo\n",
+ "mdl.fit(X_train, y_train)\n",
+ "y_predict = mdl.predict(X_test)"
+ ],
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "code",
+ "metadata": {
+ "id": "kPG5Pu2ZIn9Z",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "outputId": "b2c72156-8cc9-4115-a46f-99e637fecb64"
+ },
+ "source": [
+ "print('Classification metrics--> \\n', classification_report(y_test,y_predict))"
+ ],
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "text": [
+ "Classification metrics--> \n",
+ " precision recall f1-score support\n",
+ "\n",
+ " 0 0.98 1.00 0.99 6539\n",
+ " 1 1.00 0.98 0.99 6630\n",
+ "\n",
+ " accuracy 0.99 13169\n",
+ " macro avg 0.99 0.99 0.99 13169\n",
+ "weighted avg 0.99 0.99 0.99 13169\n",
+ "\n"
+ ],
+ "name": "stdout"
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "tCOjx9rzZFDg"
+ },
+ "source": [
+ "O resultado é ? \n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "opAJ39naYbEe"
+ },
+ "source": [
+ "Modelo |Resultado | \n",
+ "-------------------|------------------|\n",
+ "Regressão Logística | 86% \n",
+ "XGBoost | 86%\n",
+ "RandomForest | 99%"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "ID_nyzRlZiR8"
+ },
+ "source": [
+ "RandomForest ganhou novamente!!!!\n",
+ "\n",
+ "Desta vez com um alto índice de acerto 99%!\n",
+ "\n",
+ "![\"win\"](\"https://media.tenor.com/images/3a23ac35e3d051f70789f93b49001016/tenor.gif\")
\n",
+ "\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "v2IaZFwF-SG2"
+ },
+ "source": [
+ "#***Conclusão***"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "97bkoWY1e6vm"
+ },
+ "source": [
+ "![\"win\"](\"https://media3.giphy.com/media/zqaGjv9yP1RPFqw770/giphy.gif?cid=ecf05e47jnohivlqowh8s3ro1qsq5wxmot5n4okufa26d8nt&rid=giphy.gif&ct=g\")
"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "id": "MAD7OIR8bs19"
+ },
+ "source": [
+ "