Team_4_CPT.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Team_4_CPT",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Stephane-Masamba/Team_4_CPT_ML-Classification/blob/master/Team_4_CPT.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "zeuHJ4dUysfz"
      },
      "source": [
        "#                                       *Climate* *Change* *Belief* *Analysis*\n",
        "\n",
        "\n",
        "\n",
        "---\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "l6oFAvBV8aRG"
      },
      "source": [
        "#Introduction"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "5zVp58ofzaYy"
      },
      "source": [
        "Many companies are built around lessening their environmental impact and thus they offer products that are environmentally friendly and sustainable. With that, they would like to determine how people perceive climate change.\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "EQID250M2YMt"
      },
      "source": [
        "This notebook aims to build Machine Learning Classification model(s) that is able to classify accurately whether or not a person believes in climate change based on their novel tweet.\n",
        " \n",
        "\n",
        "This will be done by importing necesarry libraries as well as the training and test datasets. Data cleaning follows together with exploratory data analysis.We then wrap up the notebook by diving into different classification techniques under the Modelling section which will be followed by insights and a conclusion.\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "qVlVKIB1Kk8F"
      },
      "source": [
        "The following are some of the few classification techniques explored in this notebook:\n",
        "\n",
        "\n",
        "1.Support Vector Machines\n",
        "\n",
        "2.Naive Bayes Classfier\n",
        "\n",
        "3.K-Nearest Neighbors\n",
        "\n",
        "4.Random Forest\n",
        "\n",
        "5.Tree based models"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "4W4P-U2SJBr_"
      },
      "source": [
        "Because this is **sentiment analysis**,the interpretation and classification of emotions within text data, we will use some natural language preprocessing techniques libraries like nltk,spacy together with scikit learn."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QScBQlZF4lPa",
        "colab_type": "text"
      },
      "source": [
        "#*COMET*"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DGzzZSxP4pND",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 572
        },
        "outputId": "90357c0a-1b76-4205-8b2c-1a4d38ed1dd4"
      },
      "source": [
        "!pip install comet_ml"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting comet_ml\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/ac/d8/99564b9b0d68a8acce733a5b053497a4a51e3ffdc54e4b117808bd6c4c94/comet_ml-3.1.12-py2.py3-none-any.whl (214kB)\n",
            "\r\u001b[K     |█▌                              | 10kB 18.0MB/s eta 0:00:01\r\u001b[K     |███                             | 20kB 3.0MB/s eta 0:00:01\r\u001b[K     |████▋                           | 30kB 4.0MB/s eta 0:00:01\r\u001b[K     |██████                          | 40kB 4.3MB/s eta 0:00:01\r\u001b[K     |███████▋                        | 51kB 3.4MB/s eta 0:00:01\r\u001b[K     |█████████▏                      | 61kB 3.9MB/s eta 0:00:01\r\u001b[K     |██████████▊                     | 71kB 4.2MB/s eta 0:00:01\r\u001b[K     |████████████▏                   | 81kB 3.8MB/s eta 0:00:01\r\u001b[K     |█████████████▊                  | 92kB 4.1MB/s eta 0:00:01\r\u001b[K     |███████████████▎                | 102kB 4.4MB/s eta 0:00:01\r\u001b[K     |████████████████▊               | 112kB 4.4MB/s eta 0:00:01\r\u001b[K     |██████████████████▎             | 122kB 4.4MB/s eta 0:00:01\r\u001b[K     |███████████████████▉            | 133kB 4.4MB/s eta 0:00:01\r\u001b[K     |█████████████████████▍          | 143kB 4.4MB/s eta 0:00:01\r\u001b[K     |██████████████████████▉         | 153kB 4.4MB/s eta 0:00:01\r\u001b[K     |████████████████████████▍       | 163kB 4.4MB/s eta 0:00:01\r\u001b[K     |██████████████████████████      | 174kB 4.4MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▍    | 184kB 4.4MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████   | 194kB 4.4MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▌ | 204kB 4.4MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 215kB 4.4MB/s \n",
            "\u001b[?25hRequirement already satisfied: jsonschema!=3.1.0,>=2.6.0 in /usr/local/lib/python3.6/dist-packages (from comet_ml) (2.6.0)\n",
            "Collecting netifaces>=0.10.7\n",
            "  Downloading https://files.pythonhosted.org/packages/0c/9b/c4c7eb09189548d45939a3d3a6b3d53979c67d124459b27a094c365c347f/netifaces-0.10.9-cp36-cp36m-manylinux1_x86_64.whl\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from comet_ml) (1.12.0)\n",
            "Requirement already satisfied: requests>=2.18.4 in /usr/local/lib/python3.6/dist-packages (from comet_ml) (2.23.0)\n",
            "Collecting websocket-client>=0.55.0\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/4c/5f/f61b420143ed1c8dc69f9eaec5ff1ac36109d52c80de49d66e0c36c3dfdf/websocket_client-0.57.0-py2.py3-none-any.whl (200kB)\n",
            "\u001b[K     |████████████████████████████████| 204kB 25.1MB/s \n",
            "\u001b[?25hCollecting everett[ini]>=1.0.1; python_version >= \"3.0\"\n",
            "  Downloading https://files.pythonhosted.org/packages/12/34/de70a3d913411e40ce84966f085b5da0c6df741e28c86721114dd290aaa0/everett-1.0.2-py2.py3-none-any.whl\n",
            "Collecting comet-git-pure>=0.19.11\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/72/7a/483413046e48908986a0f9a1d8a917e1da46ae58e6ba16b2ac71b3adf8d7/comet_git_pure-0.19.16-py3-none-any.whl (409kB)\n",
            "\u001b[K     |████████████████████████████████| 419kB 24.8MB/s \n",
            "\u001b[?25hCollecting wurlitzer>=1.0.2\n",
            "  Downloading https://files.pythonhosted.org/packages/24/5e/f3bd8443bfdf96d2f5d10097d301076a9eb55637b7864e52d2d1a4d8c72a/wurlitzer-2.0.0-py2.py3-none-any.whl\n",
            "Requirement already satisfied: nvidia-ml-py3>=7.352.0 in /usr/local/lib/python3.6/dist-packages (from comet_ml) (7.352.0)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests>=2.18.4->comet_ml) (1.24.3)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests>=2.18.4->comet_ml) (2020.6.20)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests>=2.18.4->comet_ml) (3.0.4)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests>=2.18.4->comet_ml) (2.9)\n",
            "Collecting configobj; extra == \"ini\"\n",
            "  Downloading https://files.pythonhosted.org/packages/64/61/079eb60459c44929e684fa7d9e2fdca403f67d64dd9dbac27296be2e0fab/configobj-5.0.6.tar.gz\n",
            "Building wheels for collected packages: configobj\n",
            "  Building wheel for configobj (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for configobj: filename=configobj-5.0.6-cp36-none-any.whl size=34546 sha256=020e2417bb93e84679f166bc2692043ffece1268404c06aa6f8e161546cbbb05\n",
            "  Stored in directory: /root/.cache/pip/wheels/f1/e4/16/4981ca97c2d65106b49861e0b35e2660695be7219a2d351ee0\n",
            "Successfully built configobj\n",
            "Installing collected packages: netifaces, websocket-client, configobj, everett, comet-git-pure, wurlitzer, comet-ml\n",
            "Successfully installed comet-git-pure-0.19.16 comet-ml-3.1.12 configobj-5.0.6 everett-1.0.2 netifaces-0.10.9 websocket-client-0.57.0 wurlitzer-2.0.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sDLq8_is8Fra",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import comet_ml"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "agQ7LkTh0Z2q"
      },
      "source": [
        "#Imports"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "752ir0eG1BUd"
      },
      "source": [
        "###Import Libraries"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Nx6wBtTChk5U",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 398
        },
        "outputId": "d7b68372-0cbc-408a-8073-cfb1a2061d83"
      },
      "source": [
        "!pip install spacy\n",
        "!pip install NLTK"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: spacy in /usr/local/lib/python3.6/dist-packages (2.2.4)\n",
            "Requirement already satisfied: thinc==7.4.0 in /usr/local/lib/python3.6/dist-packages (from spacy) (7.4.0)\n",
            "Requirement already satisfied: preshed<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from spacy) (3.0.2)\n",
            "Requirement already satisfied: plac<1.2.0,>=0.9.6 in /usr/local/lib/python3.6/dist-packages (from spacy) (1.1.3)\n",
            "Requirement already satisfied: blis<0.5.0,>=0.4.0 in /usr/local/lib/python3.6/dist-packages (from spacy) (0.4.1)\n",
            "Requirement already satisfied: cymem<2.1.0,>=2.0.2 in /usr/local/lib/python3.6/dist-packages (from spacy) (2.0.3)\n",
            "Requirement already satisfied: catalogue<1.1.0,>=0.0.7 in /usr/local/lib/python3.6/dist-packages (from spacy) (1.0.0)\n",
            "Requirement already satisfied: numpy>=1.15.0 in /usr/local/lib/python3.6/dist-packages (from spacy) (1.18.5)\n",
            "Requirement already satisfied: tqdm<5.0.0,>=4.38.0 in /usr/local/lib/python3.6/dist-packages (from spacy) (4.41.1)\n",
            "Requirement already satisfied: requests<3.0.0,>=2.13.0 in /usr/local/lib/python3.6/dist-packages (from spacy) (2.23.0)\n",
            "Requirement already satisfied: wasabi<1.1.0,>=0.4.0 in /usr/local/lib/python3.6/dist-packages (from spacy) (0.7.0)\n",
            "Requirement already satisfied: murmurhash<1.1.0,>=0.28.0 in /usr/local/lib/python3.6/dist-packages (from spacy) (1.0.2)\n",
            "Requirement already satisfied: srsly<1.1.0,>=1.0.2 in /usr/local/lib/python3.6/dist-packages (from spacy) (1.0.2)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from spacy) (47.3.1)\n",
            "Requirement already satisfied: importlib-metadata>=0.20; python_version < \"3.8\" in /usr/local/lib/python3.6/dist-packages (from catalogue<1.1.0,>=0.0.7->spacy) (1.6.1)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests<3.0.0,>=2.13.0->spacy) (2020.6.20)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests<3.0.0,>=2.13.0->spacy) (3.0.4)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests<3.0.0,>=2.13.0->spacy) (2.9)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests<3.0.0,>=2.13.0->spacy) (1.24.3)\n",
            "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.6/dist-packages (from importlib-metadata>=0.20; python_version < \"3.8\"->catalogue<1.1.0,>=0.0.7->spacy) (3.1.0)\n",
            "Requirement already satisfied: NLTK in /usr/local/lib/python3.6/dist-packages (3.2.5)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from NLTK) (1.12.0)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "_vG_T9DtyqJe",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 141
        },
        "outputId": "7e924329-4d08-4adf-fa44-25aac83c5d7a"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import sklearn\n",
        "import matplotlib.pyplot as plt\n",
        "from numpy import arange\n",
        "import seaborn as sns\n",
        "\n",
        "#Natural Language Processing\n",
        "import re\n",
        "import spacy.cli\n",
        "import nltk\n",
        "from nltk.corpus import stopwords\n",
        "from nltk.tokenize import word_tokenize\n",
        "from sklearn.feature_extraction.text import TfidfVectorizer\n",
        "\n",
        "\n",
        "#Matrix measurement\n",
        "from sklearn.metrics import accuracy_score, f1_score\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.metrics import f1_score, precision_score, recall_score\n",
        "\n",
        "#Resampling techniques\n",
        "from sklearn.utils import resample\n",
        "from imblearn.over_sampling import SMOTE\n",
        "\n",
        "#Machine Learning Models\n",
        "from sklearn.naive_bayes import MultinomialNB, BernoulliNB\n",
        "from sklearn.svm import LinearSVC, SVC\n",
        "from sklearn.linear_model import SGDClassifier, LinearRegression\n",
        "from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier, ExtraTreesClassifier, BaggingClassifier\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "from sklearn.gaussian_process import GaussianProcessClassifier\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.model_selection import cross_val_score\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.model_selection import  KFold\n",
        "\n",
        "\n",
        "from wordcloud import WordCloud \n",
        "\n",
        "# from google.colab import drive"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
            "  import pandas.util.testing as tm\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/externals/six.py:31: FutureWarning: The module is deprecated in version 0.21 and will be removed in version 0.23 since we've dropped support for Python 2.7. Please rely on the official version of six (https://pypi.org/project/six/).\n",
            "  \"(https://pypi.org/project/six/).\", FutureWarning)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/utils/deprecation.py:144: FutureWarning: The sklearn.neighbors.base module is  deprecated in version 0.22 and will be removed in version 0.24. The corresponding classes / functions should instead be imported from sklearn.neighbors. Anything that cannot be imported from sklearn.neighbors is now part of the private API.\n",
            "  warnings.warn(message, FutureWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "m3Qrj4YjijW6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 156
        },
        "outputId": "4343a02d-a1d5-4ce6-b9b5-cde659390469"
      },
      "source": [
        "spacy.cli.download('en_core_web_sm')\n",
        "nltk.download('punkt')\n",
        "nltk.download('wordnet')\n",
        "nltk.download('stopwords')\n",
        "stop = nltk.corpus.stopwords.words('english')"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[38;5;2m✔ Download and installation successful\u001b[0m\n",
            "You can now load the model via spacy.load('en_core_web_sm')\n",
            "[nltk_data] Downloading package punkt to /root/nltk_data...\n",
            "[nltk_data]   Unzipping tokenizers/punkt.zip.\n",
            "[nltk_data] Downloading package wordnet to /root/nltk_data...\n",
            "[nltk_data]   Unzipping corpora/wordnet.zip.\n",
            "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]   Unzipping corpora/stopwords.zip.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "9wVx40yIkNQF",
        "colab": {}
      },
      "source": [
        "nlp = spacy.load('en_core_web_sm')"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "BJlh_PGlCMbF"
      },
      "source": [
        "#Read in the datasets"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "X2ZRcRRWjCn9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "c7268353-4518-48d5-dc48-86360320a75e"
      },
      "source": [
        "train = pd.read_csv('https://raw.githubusercontent.com/Stephane-Masamba/Team_4_CPT_ML-Classification/Mikael/train%20(1).csv')\n",
        "print(train.head())"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "   sentiment                                            message  tweetid\n",
            "0          1  PolySciMajor EPA chief doesn't think carbon di...   625221\n",
            "1          1  It's not like we lack evidence of anthropogeni...   126103\n",
            "2          2  RT @RawStory: Researchers say we have three ye...   698562\n",
            "3          1  #TodayinMaker# WIRED : 2016 was a pivotal year...   573736\n",
            "4          1  RT @SoyNovioDeTodas: It's 2016, and a racist, ...   466954\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Hx6btvpljPtK",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "50119a82-f8c6-4a0b-cac7-612c35f4423c"
      },
      "source": [
        "test = pd.read_csv('https://raw.githubusercontent.com/Stephane-Masamba/Team_4_CPT_ML-Classification/Mikael/test%20(1).csv')\n",
        "print(test.head())"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                                             message  tweetid\n",
            "0  Europe will now be looking to China to make su...   169760\n",
            "1  Combine this with the polling of staffers re c...    35326\n",
            "2  The scary, unimpeachable evidence that climate...   224985\n",
            "3  @Karoli @morgfair @OsborneInk @dailykos \\nPuti...   476263\n",
            "4  RT @FakeWillMoore: 'Female orgasms cause globa...   872928\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "nEX1mlx8sTik"
      },
      "source": [
        "#Data description"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "L45f3HkDsfRz"
      },
      "source": [
        "The data available aggregates tweets pertaining to climate change collected between Apr 27, 2015 and Feb 21, 2018. \n",
        "\n",
        "In total, 43943 tweets were collected. \n",
        "\n",
        "Each tweet is labelled as one of four classes.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "C1_Fc6KOtr1p"
      },
      "source": [
        "###Classes and description\n",
        "\n",
        "Class 2 : News-the tweet links to factual news about climate change\n",
        "\n",
        "Class 1 : Pro-the tweet supports the belief of man-made climate change\n",
        "\n",
        "Class 0 : Neutral-the tweet neither supports nor refutes the belief of man-made climate change\n",
        "\n",
        "Class -1 : Anti-the tweet does not believe in man-made climate change"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "2oxDAZRmsadV"
      },
      "source": [
        "###Variables definitions\n",
        "\n",
        "Sentiment: Sentiment of tweet\n",
        "\n",
        "Message: Tweet body\n",
        "\n",
        "Tweetid: Twitter unique id"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "GjJmKNotNBmv"
      },
      "source": [
        "#Data Cleaning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "zwefjbFg2T1L"
      },
      "source": [
        "Knowing that we are dealing with text data, we decided to first clean the data by making all tweets lower-case, removing punctuation marks and removing white spaces before doing anything else. Also, replacing all links with the word 'LINK' and all user handles with 'USER_REF'"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "-FT0mtEVjZv2",
        "colab": {}
      },
      "source": [
        "def clean_text(df):\n",
        "  i = 0\n",
        "  for tweet in df['message']:\n",
        "    tweet = tweet.lower()\n",
        "    tweet = re.sub(r'http\\S+', 'LINK', tweet)\n",
        "    tweet = re.sub(r'@\\S+', 'USER_REF', tweet)\n",
        "    tweet = re.sub(r'[^\\w\\s]', '', tweet)\n",
        "    tweet = tweet.lstrip()\n",
        "    tweet = tweet.rstrip()\n",
        "    tweet = tweet.replace('  ', ' ')\n",
        "    df.loc[i, 'message'] = tweet\n",
        "    i += 1\n",
        "\n"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "imNo7YM9PDvu",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "ad846c8d-c545-4cf3-811a-d3c303542703"
      },
      "source": [
        "clean_text(train)\n",
        "train"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>sentiment</th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>polyscimajor epa chief doesnt think carbon dio...</td>\n",
              "      <td>625221</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>its not like we lack evidence of anthropogenic...</td>\n",
              "      <td>126103</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>rt USER_REF researchers say we have three year...</td>\n",
              "      <td>698562</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1</td>\n",
              "      <td>todayinmaker wired 2016 was a pivotal year in ...</td>\n",
              "      <td>573736</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1</td>\n",
              "      <td>rt USER_REF its 2016 and a racist sexist clima...</td>\n",
              "      <td>466954</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15814</th>\n",
              "      <td>1</td>\n",
              "      <td>rt USER_REF they took down the material on glo...</td>\n",
              "      <td>22001</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15815</th>\n",
              "      <td>2</td>\n",
              "      <td>rt USER_REF how climate change could be breaki...</td>\n",
              "      <td>17856</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15816</th>\n",
              "      <td>0</td>\n",
              "      <td>notiven rt nytimesworld what does trump actual...</td>\n",
              "      <td>384248</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15817</th>\n",
              "      <td>-1</td>\n",
              "      <td>rt USER_REF hey liberals the climate change cr...</td>\n",
              "      <td>819732</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15818</th>\n",
              "      <td>0</td>\n",
              "      <td>rt USER_REF USER_REF climate change equation i...</td>\n",
              "      <td>806319</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>15819 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       sentiment                                            message  tweetid\n",
              "0              1  polyscimajor epa chief doesnt think carbon dio...   625221\n",
              "1              1  its not like we lack evidence of anthropogenic...   126103\n",
              "2              2  rt USER_REF researchers say we have three year...   698562\n",
              "3              1  todayinmaker wired 2016 was a pivotal year in ...   573736\n",
              "4              1  rt USER_REF its 2016 and a racist sexist clima...   466954\n",
              "...          ...                                                ...      ...\n",
              "15814          1  rt USER_REF they took down the material on glo...    22001\n",
              "15815          2  rt USER_REF how climate change could be breaki...    17856\n",
              "15816          0  notiven rt nytimesworld what does trump actual...   384248\n",
              "15817         -1  rt USER_REF hey liberals the climate change cr...   819732\n",
              "15818          0  rt USER_REF USER_REF climate change equation i...   806319\n",
              "\n",
              "[15819 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "XZGhAPi_TIl6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "52ef4fc4-90e3-4888-d32b-9c271d9982e5"
      },
      "source": [
        "clean_text(test)\n",
        "test"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>europe will now be looking to china to make su...</td>\n",
              "      <td>169760</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>combine this with the polling of staffers re c...</td>\n",
              "      <td>35326</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>the scary unimpeachable evidence that climate ...</td>\n",
              "      <td>224985</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>USER_REF USER_REF USER_REF USER_REF \\nputin go...</td>\n",
              "      <td>476263</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>rt USER_REF female orgasms cause global warmin...</td>\n",
              "      <td>872928</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10541</th>\n",
              "      <td>rt USER_REF brb writing a poem about climate c...</td>\n",
              "      <td>895714</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10542</th>\n",
              "      <td>2016 the year climate change came home during ...</td>\n",
              "      <td>875167</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10543</th>\n",
              "      <td>rt USER_REF pacific countries positive about f...</td>\n",
              "      <td>78329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10544</th>\n",
              "      <td>rt USER_REF youre so hot you must be the cause...</td>\n",
              "      <td>867455</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10545</th>\n",
              "      <td>rt USER_REF climate change is a global issue t...</td>\n",
              "      <td>470892</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>10546 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                 message  tweetid\n",
              "0      europe will now be looking to china to make su...   169760\n",
              "1      combine this with the polling of staffers re c...    35326\n",
              "2      the scary unimpeachable evidence that climate ...   224985\n",
              "3      USER_REF USER_REF USER_REF USER_REF \\nputin go...   476263\n",
              "4      rt USER_REF female orgasms cause global warmin...   872928\n",
              "...                                                  ...      ...\n",
              "10541  rt USER_REF brb writing a poem about climate c...   895714\n",
              "10542  2016 the year climate change came home during ...   875167\n",
              "10543  rt USER_REF pacific countries positive about f...    78329\n",
              "10544  rt USER_REF youre so hot you must be the cause...   867455\n",
              "10545  rt USER_REF climate change is a global issue t...   470892\n",
              "\n",
              "[10546 rows x 2 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "IGiWpv04215t"
      },
      "source": [
        "Here we add 'LINK' and 'USER_REF' to our list of stop words and remove all stop words from each tweet"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "bRaM34T3jkNJ",
        "colab": {}
      },
      "source": [
        "def remove_stopwords(df):\n",
        "    my_stop_words = stopwords.words('english')\n",
        "    my_stop_words.append('LINK')\n",
        "    my_stop_words.append('USER_REF')\n",
        "\n",
        "    df_index = 0\n",
        "\n",
        "    for tweet in df['message']:\n",
        "      tweet = word_tokenize(tweet)\n",
        "      tweet = [word for word in tweet if not word in my_stop_words]\n",
        "      tweet = ' '.join(tweet)\n",
        "\n",
        "      df.loc[df_index, 'message'] = tweet\n",
        "      df_index += 1\n",
        "\n",
        "    "
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "-k21QOvEBoUi",
        "colab": {}
      },
      "source": [
        "remove_stopwords(train)"
      ],
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xyE2jaErW7Wt",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "3157eac7-353f-4b2e-990b-5d70db54f3b2"
      },
      "source": [
        "train"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>sentiment</th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>ORG chief donot think carbon dioxide main caus...</td>\n",
              "      <td>625221</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>like lack evidence anthropogenic global warming</td>\n",
              "      <td>126103</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>rt researchers say three years act climate cha...</td>\n",
              "      <td>698562</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1</td>\n",
              "      <td>todayinmaker wired 2016 pivotal year war clima...</td>\n",
              "      <td>573736</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1</td>\n",
              "      <td>rt 2016 racist sexist climate change denying b...</td>\n",
              "      <td>466954</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15814</th>\n",
              "      <td>1</td>\n",
              "      <td>rt took material global warming lgbt rights he...</td>\n",
              "      <td>22001</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15815</th>\n",
              "      <td>2</td>\n",
              "      <td>rt climate change could breaking 200millionyea...</td>\n",
              "      <td>17856</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15816</th>\n",
              "      <td>0</td>\n",
              "      <td>notiven rt nytimesworld trump actually believe...</td>\n",
              "      <td>384248</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15817</th>\n",
              "      <td>-1</td>\n",
              "      <td>rt hey liberals climate change crap hoax ties ...</td>\n",
              "      <td>819732</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15818</th>\n",
              "      <td>0</td>\n",
              "      <td>rt climate change equation 4 screenshots</td>\n",
              "      <td>806319</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>15819 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       sentiment                                            message  tweetid\n",
              "0              1  ORG chief donot think carbon dioxide main caus...   625221\n",
              "1              1    like lack evidence anthropogenic global warming   126103\n",
              "2              2  rt researchers say three years act climate cha...   698562\n",
              "3              1  todayinmaker wired 2016 pivotal year war clima...   573736\n",
              "4              1  rt 2016 racist sexist climate change denying b...   466954\n",
              "...          ...                                                ...      ...\n",
              "15814          1  rt took material global warming lgbt rights he...    22001\n",
              "15815          2  rt climate change could breaking 200millionyea...    17856\n",
              "15816          0  notiven rt nytimesworld trump actually believe...   384248\n",
              "15817         -1  rt hey liberals climate change crap hoax ties ...   819732\n",
              "15818          0           rt climate change equation 4 screenshots   806319\n",
              "\n",
              "[15819 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "E2Szdx35TM4F",
        "colab": {}
      },
      "source": [
        "remove_stopwords(test)"
      ],
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u1qElPPGW-hz",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "62588d2d-cae7-4042-fe78-5c6743ccd2d0"
      },
      "source": [
        "test"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>LOC look gpe make sure alone fight climate change</td>\n",
              "      <td>169760</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>combine polling staffers climate change womens...</td>\n",
              "      <td>35326</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>scary unimpeachable evidence climate change al...</td>\n",
              "      <td>224985</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>putin got jill trump doesnt believe climate ch...</td>\n",
              "      <td>476263</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>rt female orgasms cause global warming sarcast...</td>\n",
              "      <td>872928</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10541</th>\n",
              "      <td>rt brb writing poem climate change climatechan...</td>\n",
              "      <td>895714</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10542</th>\n",
              "      <td>2016 year climate change came home hottest yea...</td>\n",
              "      <td>875167</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10543</th>\n",
              "      <td>rt pacific countries positive fiji leading glo...</td>\n",
              "      <td>78329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10544</th>\n",
              "      <td>rt youre hot must cause global warming aldubla...</td>\n",
              "      <td>867455</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10545</th>\n",
              "      <td>rt climate change global issue thats getting w...</td>\n",
              "      <td>470892</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>10546 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                 message  tweetid\n",
              "0      LOC look gpe make sure alone fight climate change   169760\n",
              "1      combine polling staffers climate change womens...    35326\n",
              "2      scary unimpeachable evidence climate change al...   224985\n",
              "3      putin got jill trump doesnt believe climate ch...   476263\n",
              "4      rt female orgasms cause global warming sarcast...   872928\n",
              "...                                                  ...      ...\n",
              "10541  rt brb writing poem climate change climatechan...   895714\n",
              "10542  2016 year climate change came home hottest yea...   875167\n",
              "10543  rt pacific countries positive fiji leading glo...    78329\n",
              "10544  rt youre hot must cause global warming aldubla...   867455\n",
              "10545  rt climate change global issue thats getting w...   470892\n",
              "\n",
              "[10546 rows x 2 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "jsRm6MAD3Gqb"
      },
      "source": [
        "###Find all named entities for each tweet"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "4HISNTJcj9tI",
        "colab": {}
      },
      "source": [
        "def entities(df):\n",
        "    df_index = 0\n",
        "\n",
        "    for tweet in df['message']:\n",
        "      tweet = nlp(tweet)\n",
        "\n",
        "      for entity in tweet.ents:\n",
        "        df.loc[df_index, 'message'] = df.loc[df_index, 'message'].replace(str(entity.text), str(entity.label_))\n",
        "\n",
        "      df_index += 1\n",
        "\n",
        "      "
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "ohas__j-QFGV",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "c6c3830a-9ef9-4b27-a683-ebd30a6e7cbb"
      },
      "source": [
        "entities(train)\n",
        "train"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>sentiment</th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>ORG chief donot think carbon dioxide main caus...</td>\n",
              "      <td>625221</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>like lack evidence anthropogenic global warming</td>\n",
              "      <td>126103</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>rt researchers say DATE act climate change late</td>\n",
              "      <td>698562</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1</td>\n",
              "      <td>todayinmaker wired DATE pivotal year war clima...</td>\n",
              "      <td>573736</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1</td>\n",
              "      <td>rt DATE racist sexist climate change denying b...</td>\n",
              "      <td>466954</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15814</th>\n",
              "      <td>1</td>\n",
              "      <td>rt took material global warming lgbt rights he...</td>\n",
              "      <td>22001</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15815</th>\n",
              "      <td>2</td>\n",
              "      <td>rt climate change could breaking CARDINAL rela...</td>\n",
              "      <td>17856</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15816</th>\n",
              "      <td>0</td>\n",
              "      <td>PERSON rt nytimesworld trump actually believe ...</td>\n",
              "      <td>384248</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15817</th>\n",
              "      <td>-1</td>\n",
              "      <td>rt hey liberals climate change crap hoax ties ...</td>\n",
              "      <td>819732</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15818</th>\n",
              "      <td>0</td>\n",
              "      <td>rt climate change equation CARDINAL screenshots</td>\n",
              "      <td>806319</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>15819 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       sentiment                                            message  tweetid\n",
              "0              1  ORG chief donot think carbon dioxide main caus...   625221\n",
              "1              1    like lack evidence anthropogenic global warming   126103\n",
              "2              2    rt researchers say DATE act climate change late   698562\n",
              "3              1  todayinmaker wired DATE pivotal year war clima...   573736\n",
              "4              1  rt DATE racist sexist climate change denying b...   466954\n",
              "...          ...                                                ...      ...\n",
              "15814          1  rt took material global warming lgbt rights he...    22001\n",
              "15815          2  rt climate change could breaking CARDINAL rela...    17856\n",
              "15816          0  PERSON rt nytimesworld trump actually believe ...   384248\n",
              "15817         -1  rt hey liberals climate change crap hoax ties ...   819732\n",
              "15818          0    rt climate change equation CARDINAL screenshots   806319\n",
              "\n",
              "[15819 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "lmL3Ad3rTZHC",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "4d475f85-edb4-46b8-bc6a-d5174eac5c97"
      },
      "source": [
        "entities(test)\n",
        "test"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>LOC look gpe make sure alone fight climate change</td>\n",
              "      <td>169760</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>combine polling staffers climate change womens...</td>\n",
              "      <td>35326</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>scary unimpeachable evidence climate change al...</td>\n",
              "      <td>224985</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>PERSON got PERSON doesnt believe climate chang...</td>\n",
              "      <td>476263</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>rt female orgasms cause global warming sarcast...</td>\n",
              "      <td>872928</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10541</th>\n",
              "      <td>rt brb writing poem climate change climatechan...</td>\n",
              "      <td>895714</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10542</th>\n",
              "      <td>DATE climate change came home hottest DATE rec...</td>\n",
              "      <td>875167</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10543</th>\n",
              "      <td>rt pacific countries positive GPE leading glob...</td>\n",
              "      <td>78329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10544</th>\n",
              "      <td>rt youre hot must cause global warming aldubla...</td>\n",
              "      <td>867455</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10545</th>\n",
              "      <td>rt climate change global issue thats getting w...</td>\n",
              "      <td>470892</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>10546 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                 message  tweetid\n",
              "0      LOC look gpe make sure alone fight climate change   169760\n",
              "1      combine polling staffers climate change womens...    35326\n",
              "2      scary unimpeachable evidence climate change al...   224985\n",
              "3      PERSON got PERSON doesnt believe climate chang...   476263\n",
              "4      rt female orgasms cause global warming sarcast...   872928\n",
              "...                                                  ...      ...\n",
              "10541  rt brb writing poem climate change climatechan...   895714\n",
              "10542  DATE climate change came home hottest DATE rec...   875167\n",
              "10543  rt pacific countries positive GPE leading glob...    78329\n",
              "10544  rt youre hot must cause global warming aldubla...   867455\n",
              "10545  rt climate change global issue thats getting w...   470892\n",
              "\n",
              "[10546 rows x 2 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "tCbgGYy83POz"
      },
      "source": [
        "###Lemmatize every word in each tweet"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "A490Wi48k_gs",
        "colab": {}
      },
      "source": [
        "def lem_text(df):\n",
        "    df_index = 0\n",
        "\n",
        "    for tweet in df['message']:\n",
        "      tweet = nlp(tweet)\n",
        "      \n",
        "      for token in tweet:\n",
        "        df.loc[df_index, 'message'] = df.loc[df_index, 'message'].replace(str(token.text), str(token.lemma_))\n",
        "\n",
        "      df_index += 1\n",
        "\n",
        "      "
      ],
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "LNMW3AsWRVAA",
        "colab": {}
      },
      "source": [
        "lem_text(train)"
      ],
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7ZulIwtdYMpH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "7bd8e901-cde6-4489-b90e-3adecdcc2763"
      },
      "source": [
        "train"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>sentiment</th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>ORG chief donot think carbon dioxide main caus...</td>\n",
              "      <td>625221</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>like lack evidence anthropogenic global warming</td>\n",
              "      <td>126103</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>rt researcher say DATE act climate change late</td>\n",
              "      <td>698562</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1</td>\n",
              "      <td>todayinmaker wire DATE pivotal year war climat...</td>\n",
              "      <td>573736</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1</td>\n",
              "      <td>rt DATE racist sexist climate change deny bigo...</td>\n",
              "      <td>466954</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15814</th>\n",
              "      <td>1</td>\n",
              "      <td>rt take material global warming lgbt right hea...</td>\n",
              "      <td>22001</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15815</th>\n",
              "      <td>2</td>\n",
              "      <td>rt climate change could break CARDINAL relatio...</td>\n",
              "      <td>17856</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15816</th>\n",
              "      <td>0</td>\n",
              "      <td>person rt nytimesworld trump actually believe ...</td>\n",
              "      <td>384248</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15817</th>\n",
              "      <td>-1</td>\n",
              "      <td>rt hey liberal climate change crap hoax ties a...</td>\n",
              "      <td>819732</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15818</th>\n",
              "      <td>0</td>\n",
              "      <td>rt climate change equation CARDINAL screenshot</td>\n",
              "      <td>806319</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>15819 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       sentiment                                            message  tweetid\n",
              "0              1  ORG chief donot think carbon dioxide main caus...   625221\n",
              "1              1    like lack evidence anthropogenic global warming   126103\n",
              "2              2     rt researcher say DATE act climate change late   698562\n",
              "3              1  todayinmaker wire DATE pivotal year war climat...   573736\n",
              "4              1  rt DATE racist sexist climate change deny bigo...   466954\n",
              "...          ...                                                ...      ...\n",
              "15814          1  rt take material global warming lgbt right hea...    22001\n",
              "15815          2  rt climate change could break CARDINAL relatio...    17856\n",
              "15816          0  person rt nytimesworld trump actually believe ...   384248\n",
              "15817         -1  rt hey liberal climate change crap hoax ties a...   819732\n",
              "15818          0     rt climate change equation CARDINAL screenshot   806319\n",
              "\n",
              "[15819 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "7cj6gDObTccc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "651c7fb2-6ca8-42b3-df93-76bad322b75d"
      },
      "source": [
        "lem_text(test)\n",
        "test"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>LOC look gpe make sure alone fight climate change</td>\n",
              "      <td>169760</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>combine polling staffer climate change women r...</td>\n",
              "      <td>35326</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>scary unimpeachable evidence climate change al...</td>\n",
              "      <td>224985</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>person get person donot believe climate change...</td>\n",
              "      <td>476263</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>rt female orgasm cause global warm sarcastic NORP</td>\n",
              "      <td>872928</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10541</th>\n",
              "      <td>rt brb write poem climate change climatechange...</td>\n",
              "      <td>895714</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10542</th>\n",
              "      <td>date climate change come home hot date record ...</td>\n",
              "      <td>875167</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10543</th>\n",
              "      <td>rt pacific country positive GPE lead global cl...</td>\n",
              "      <td>78329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10544</th>\n",
              "      <td>rt -PRON-be hot must cause global warming aldu...</td>\n",
              "      <td>867455</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10545</th>\n",
              "      <td>rt climate change global issue thats get bad e...</td>\n",
              "      <td>470892</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>10546 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                 message  tweetid\n",
              "0      LOC look gpe make sure alone fight climate change   169760\n",
              "1      combine polling staffer climate change women r...    35326\n",
              "2      scary unimpeachable evidence climate change al...   224985\n",
              "3      person get person donot believe climate change...   476263\n",
              "4      rt female orgasm cause global warm sarcastic NORP   872928\n",
              "...                                                  ...      ...\n",
              "10541  rt brb write poem climate change climatechange...   895714\n",
              "10542  date climate change come home hot date record ...   875167\n",
              "10543  rt pacific country positive GPE lead global cl...    78329\n",
              "10544  rt -PRON-be hot must cause global warming aldu...   867455\n",
              "10545  rt climate change global issue thats get bad e...   470892\n",
              "\n",
              "[10546 rows x 2 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "VjfJWKpqUDbH"
      },
      "source": [
        "#Exploratory Data Analysis"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "ntJYpsbHViGN"
      },
      "source": [
        "Now it is time to analyse the data and how it is structured."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "qjV-NA1yXgo8"
      },
      "source": [
        "###Missing values"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "TdgDkv4rWWtF",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        },
        "outputId": "3d3527c0-2fcd-4aee-f598-ca7d38edeae2"
      },
      "source": [
        "train.isnull().sum()\n"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "sentiment    0\n",
              "message      0\n",
              "tweetid      0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "qn7bl0_jWJMk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "outputId": "4e504ad8-4e8a-42bf-df11-f8199d9233c6"
      },
      "source": [
        "test.isnull().sum()"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "message    0\n",
              "tweetid    0\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "3-2YhdWmWu_h"
      },
      "source": [
        "There are no missing values in both the train and test datasets."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "8EzkGsU9XoOu"
      },
      "source": [
        "###Sentiment Classes"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "dDrbGvPtW2sz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        },
        "outputId": "b4fb2467-f2d1-4c5a-eed3-3a8c62ddc47e"
      },
      "source": [
        "train.sentiment.value_counts()"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              " 1    8530\n",
              " 2    3640\n",
              " 0    2353\n",
              "-1    1296\n",
              "Name: sentiment, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "CrTxP63uXCAW",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 278
        },
        "outputId": "e4861b0f-37df-46cb-a4cf-df4907f0a040"
      },
      "source": [
        "counts = train[\"sentiment\"].value_counts()\n",
        "plt.bar(range(len(counts)), counts)\n",
        "plt.xticks([0, 1, 2, 3], ['Pro', 'News', 'Neutral', 'Anti'])\n",
        "\n",
        "plt.ylabel(\"Total per class\")\n",
        "plt.xlabel(\"Sentiment Classes\")\n",
        "plt.show()"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "gPMW7s8enuWc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "f969c671-75f0-4903-b256-87ba99470115"
      },
      "source": [
        "#Percentage of the major class\n",
        "len(train[train.sentiment==1])/len(train.sentiment)"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.5392249826158417"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "e8IBeBUJXvo4"
      },
      "source": [
        "The 'Pro' climate change class accounts for 54% of the data set and the remaining 46% is shared amongst the three other classes. This imbalance indicates how biased our model might me thus exploring resampling techniques might be required in order to improve the model accuracy."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QB2WJOjW2ZDx",
        "colab_type": "text"
      },
      "source": [
        "###Word Cloud"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mAvHt7I82bhL",
        "colab_type": "text"
      },
      "source": [
        "Below we show some of the words used in the different tweets under each class."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "mUjTqUPtYlXF",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 744
        },
        "outputId": "6085019b-9ffb-4924-b5a7-b6f1e9e3d7ff"
      },
      "source": [
        "#word clouds\n",
        "news = train[train['sentiment'] == 2]['message']\n",
        "pro = train[train['sentiment'] == 1]['message']\n",
        "neutral =train[train['sentiment'] == 0]['message']\n",
        "Anti = train[train['sentiment'] ==-1]['message']\n",
        "\n",
        "\n",
        "news = [word for line in news for word in line.split()]\n",
        "pro = [word for line in pro for word in line.split()]\n",
        "neutral = [word for line in neutral for word in line.split()]\n",
        "Anti= [word for line in Anti for word in line.split()]\n",
        "\n",
        "news = WordCloud(\n",
        "    background_color='white',\n",
        "    max_words=20,\n",
        "    max_font_size=40,\n",
        "    scale=5,\n",
        "    random_state=1,\n",
        "    collocations=False,\n",
        "    normalize_plurals=False\n",
        ").generate(' '.join(news))\n",
        "\n",
        "pro = WordCloud(\n",
        "    background_color='white',\n",
        "    max_words=20,\n",
        "    max_font_size=40,\n",
        "    scale=5,\n",
        "    random_state=1,\n",
        "    collocations=False,\n",
        "    normalize_plurals=False\n",
        ").generate(' '.join(pro))\n",
        "\n",
        "\n",
        "\n",
        "neutral = WordCloud(\n",
        "    background_color='white',\n",
        "    max_words=20,\n",
        "    max_font_size=40,\n",
        "    scale=5,\n",
        "    random_state=1,\n",
        "    collocations=False,\n",
        "    normalize_plurals=False\n",
        ").generate(' '.join(neutral))\n",
        "\n",
        "\n",
        "Anti = WordCloud(\n",
        "    background_color='white',\n",
        "    max_words=20,\n",
        "    max_font_size=40,\n",
        "    scale=5,\n",
        "    random_state=1,\n",
        "    collocations=False,\n",
        "    normalize_plurals=False\n",
        ").generate(' '.join(Anti))\n",
        "\n",
        "\n",
        "fig, axs = plt.subplots(2, 2, figsize = (20, 10))\n",
        "fig.tight_layout(pad = 0)\n",
        "\n",
        "axs[0, 0].imshow(news)\n",
        "axs[0, 0].set_title('Words from news tweets', fontsize = 20)\n",
        "axs[0, 0].axis('off')\n",
        "\n",
        "axs[0, 1].imshow(pro)\n",
        "axs[0, 1].set_title('Words from pro tweets', fontsize = 20)\n",
        "axs[0, 1].axis('off')\n",
        "\n",
        "\n",
        "axs[1, 0].imshow(Anti)\n",
        "axs[1, 0].set_title('Words from anti tweets', fontsize = 20)\n",
        "axs[1, 0].axis('off')\n",
        "\n",
        "axs[1, 1].imshow(neutral)\n",
        "axs[1, 1].set_title('Words from neutral tweets', fontsize = 20)\n",
        "axs[1, 1].axis('off')\n",
        "\n",
        "plt.savefig('joint_cloud.png')"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x720 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "OdmIKBZQ3lnv"
      },
      "source": [
        "#Define our feature and target variable to use when modelling"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "STTboX79ozzr",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "outputId": "bda7a6eb-cc0f-46fb-d95d-208b3fb0483c"
      },
      "source": [
        "X = train['message']\n",
        "X"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0        ORG chief donot think carbon dioxide main caus...\n",
              "1          like lack evidence anthropogenic global warming\n",
              "2           rt researcher say DATE act climate change late\n",
              "3        todayinmaker wire DATE pivotal year war climat...\n",
              "4        rt DATE racist sexist climate change deny bigo...\n",
              "                               ...                        \n",
              "15814    rt take material global warming lgbt right hea...\n",
              "15815    rt climate change could break CARDINAL relatio...\n",
              "15816    person rt nytimesworld trump actually believe ...\n",
              "15817    rt hey liberal climate change crap hoax ties a...\n",
              "15818       rt climate change equation CARDINAL screenshot\n",
              "Name: message, Length: 15819, dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "S9v3Daf7o5VN",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "outputId": "6676b5a8-75d8-4d9a-93d7-517e89a70039"
      },
      "source": [
        "y = train['sentiment']\n",
        "y"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0        1\n",
              "1        1\n",
              "2        2\n",
              "3        1\n",
              "4        1\n",
              "        ..\n",
              "15814    1\n",
              "15815    2\n",
              "15816    0\n",
              "15817   -1\n",
              "15818    0\n",
              "Name: sentiment, Length: 15819, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "z48y9g7HaDH_"
      },
      "source": [
        "###Convert features to machine language"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "EdJlKYIm3ELO",
        "colab": {}
      },
      "source": [
        "tf_vecto = TfidfVectorizer(lowercase = True,stop_words = 'english',ngram_range=(1, 2))\n",
        "X = tf_vecto.fit_transform(X)"
      ],
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "u-LT4oVwZubv"
      },
      "source": [
        "###Splitting the data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "vm_kO-V5Z0n2",
        "colab": {}
      },
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.2, random_state = 42)"
      ],
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "3EcLpNelpOj-",
        "colab": {}
      },
      "source": [
        "#X_test = test['message']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "13mORVJNpRZ0",
        "colab": {}
      },
      "source": [
        "#X_test = tf_vect.transform(X_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "vyhuwyLh3oMX"
      },
      "source": [
        "#Modelling"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "ck4qTN4AkIee"
      },
      "source": [
        "##Modelling with maximum data cleaning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "is1L3SPwa1x-"
      },
      "source": [
        "The function below takes in a model and splits our training data into a train and test set. Once this is done, the function calculates the accuracy score for both the train and test set and returns the avergace accuracy score for both."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "0slY90EHJ3y5",
        "colab": {}
      },
      "source": [
        "def accuracy(model):\n",
        "    features = train['message']\n",
        "    target = train['sentiment']\n",
        "\n",
        "    train_scores = []\n",
        "    test_scores = []\n",
        "\n",
        "    #tf_vect = TfidfVectorizer(ngram_range=(1, 2))\n",
        "    tf_vecto = TfidfVectorizer(lowercase = True,stop_words = 'english',ngram_range=(1, 2))\n",
        "\n",
        "\n",
        "    folds = KFold(n_splits=5, shuffle=True)\n",
        "\n",
        "    for train_index, test_index in folds.split(features):\n",
        "        x_train, x_test = features.iloc[train_index], features.iloc[test_index]    \n",
        "        y_train, y_test = target.iloc[train_index], target.iloc[test_index]\n",
        "\n",
        "        x_train = tf_vecto.fit_transform(x_train)\n",
        "        x_test = tf_vecto.transform(x_test)\n",
        "     \n",
        "        model.fit(x_train, y_train)\n",
        "        train_predictions = model.predict(x_train)\n",
        "        test_predictions = model.predict(x_test)\n",
        "\n",
        "        train_score = accuracy_score(y_train, train_predictions)\n",
        "        train_scores.append(train_score)\n",
        "\n",
        "        test_score = accuracy_score(y_test, test_predictions)\n",
        "        test_scores.append(test_score)\n",
        "\n",
        "    avg_train_accuracy = np.mean(train_scores)\n",
        "    avg_test_accuracy = np.mean(test_scores)\n",
        "\n",
        "    return [avg_train_accuracy, avg_test_accuracy]\n"
      ],
      "execution_count": 45,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "LMRZCX07bRVv"
      },
      "source": [
        "###Support Vector Machines(SVM)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "r_X5tg35bX9Z"
      },
      "source": [
        "SVM or Support Vector Machine is a linear model for classification and regression problems. It can solve linear and non-linear problems and work well for many practical problems. The idea of SVM is simple: The algorithm creates a line or a hyperplane which separates the data into classes"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "D4wFMtmIpU5R",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "d1293e3d-e157-4b07-c9ac-6735f1431d39"
      },
      "source": [
        "sv = SVC()\n",
        "\n",
        "sv_accuracy = accuracy(sv)\n",
        "sv_accuracy"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9639831962302094, 0.7025728906409565]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "2ZjxnRYBfHQE"
      },
      "source": [
        "###Bernoulli Naive Bayes"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "SvqbdCtIFdsn"
      },
      "source": [
        "The Bernoulli naive Bayes classifier assumes that all our features are binary such that they take only two values (e.g. a nominal categorical feature that has been one-hot encoded)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Wf9-alNPprIm",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "ad90e641-a972-47f1-e6f4-c172892b6b64"
      },
      "source": [
        "bernoulli = BernoulliNB()\n",
        "\n",
        "bernoulli_accuracy = accuracy(bernoulli)\n",
        "bernoulli_accuracy"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.7294708884172543, 0.618686911280198]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 47
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "n_OAlcHYtBB-"
      },
      "source": [
        "###Multi Nomial Naive Bayes"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "plXkTQxSFsBe"
      },
      "source": [
        "In summary, Naive Bayes classifier is a general term which refers to conditional independence of each of the features in the model, while Multinomial Naive Bayes classifier is a specific instance of a Naive Bayes classifier which uses a multinomial distribution for each of the features"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "3esjXnzMpyTZ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "8d17826c-065d-4b8e-96b8-b19ebe441364"
      },
      "source": [
        "mnb = MultinomialNB()\n",
        "\n",
        "mnb_accuracy = accuracy(mnb)\n",
        "mnb_accuracy"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.7008660186381661, 0.6268405468891454]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "jwsTSgGZtHAQ"
      },
      "source": [
        "###Stochastic Gradient Descent (SGD)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PRF4T8IdplQO",
        "colab_type": "text"
      },
      "source": [
        " SGD is a simple,efficient approach to fitting linear classifiers and regressors under convex loss functions such as (linear) Support Vector Machines and Logistic Regression.It has received a considerable amount of attention just recently in the context of large-scale learning."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "lh0reHXQp5sK",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "b2d45551-0d4f-4caa-d681-5cb17b4965f0"
      },
      "source": [
        "sgd = SGDClassifier()\n",
        "\n",
        "sgd_accuracy = accuracy(sgd)\n",
        "sgd_accuracy"
      ],
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9348884264950268, 0.7162905441512623]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "JH-K4IHUcAGT"
      },
      "source": [
        "###Random Forest"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "js51Mz9CF_e9"
      },
      "source": [
        "The random forest is a classification algorithm consisting of many decisions trees. It uses bagging and feature randomness when building each individual tree to try to create an uncorrelated forest of trees whose prediction by committee is more accurate than that of any individual tree."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "kmzq-NeJqA4V",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "35e7e8e1-a429-4e47-8a7e-6cffb0a022a2"
      },
      "source": [
        "rand_forest = RandomForestClassifier()\n",
        "\n",
        "rand_forest_accuracy = accuracy(rand_forest)\n",
        "rand_forest_accuracy"
      ],
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9963651355305464, 0.6726722698009899]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "UjFVP6SycIY3"
      },
      "source": [
        "###K-Nearest Neighbors (KNN)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "KxiZtmi4GCwF"
      },
      "source": [
        "In pattern recognition, the k-nearest neighbors algorithm (k-NN) is a non-parametric method proposed by Thomas Cover used for classification and regression. In both cases, the input consists of the k closest training examples in the feature space. ... In k-NN classification, the output is a class membership."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "n0otZVnWqILq",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "e21d2d4a-0996-47d2-ed5e-1b0b3b8ca5f0"
      },
      "source": [
        "knn = KNeighborsClassifier()\n",
        "\n",
        "knn_accuracy = accuracy(knn)\n",
        "knn_accuracy"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.7265945936631033, 0.6058532342792553]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "8e5t7CvUtL9r"
      },
      "source": [
        "###Gradient Boosting Classifier"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "ccOol_1mGJ3h"
      },
      "source": [
        "Gradient boosting is a machine learning technique for regression and classification problems, which produces a prediction model in the form of an ensemble of weak prediction models, typically decision trees."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "k0sylfPpqO_t",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "8693334f-bdc1-4fae-e3df-8699ab7b0229"
      },
      "source": [
        "grad_booster = GradientBoostingClassifier()\n",
        "\n",
        "grad_booster_accuracy = accuracy(grad_booster)\n",
        "grad_booster_accuracy"
      ],
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.6870377570964541, 0.6411277999850515]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 53
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "TiY-u-86tSJ9"
      },
      "source": [
        "###Extra Trees Classifier"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "K1VCsoUTGNw2"
      },
      "source": [
        "An extra-trees classifier is a class that implements a meta estimator that fits a number of randomized decision trees (a.k.a. extra-trees) on various sub-samples of the dataset and uses averaging to improve the predictive accuracy and control over-fitting. The number of trees in the forest."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "KZclsZ0sqYtk",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "1bb6cc26-916d-4c9e-ade1-a1e81c8b2431"
      },
      "source": [
        "extra_trees = ExtraTreesClassifier()\n",
        "\n",
        "extra_trees_accuracy = accuracy(extra_trees)\n",
        "extra_trees_accuracy"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9963335149831096, 0.6960607658158712]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "xgGyoTzntZJR"
      },
      "source": [
        "###Bagging"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "wUv9fCS4GRD_"
      },
      "source": [
        "A Bagging classifier is an ensemble meta-estimator that fits base classifiers each on random subsets of the original dataset and then aggregate their individual predictions (either by voting or by averaging) to form a final prediction. ... If None, then the base estimator is a decision tree."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "1tS8Jwndqhql",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "39c5f62a-3397-4039-f0fc-e9a4fef6e5c5"
      },
      "source": [
        "bagging = BaggingClassifier()\n",
        "\n",
        "bagging_accuracy = accuracy(bagging)\n",
        "bagging_accuracy"
      ],
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9728332782223562, 0.6158413914361416]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 55
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "hjgZrHEbtcD0"
      },
      "source": [
        "###Decision Tree Classifier"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "KISGcMJPGUxC"
      },
      "source": [
        "The decision tree classifier (Pang-Ning et al., 2006) creates the classification model by building a decision tree. Each node in the tree specifies a test on an attribute, each branch descending from that node corresponds to one of the possible values for that attribute."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "QfGgsZOCqmaR",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "bce0925a-9aca-4cab-d874-557d6a306a31"
      },
      "source": [
        "dec_tree = DecisionTreeClassifier()\n",
        "\n",
        "dec_tree_accuracy = accuracy(dec_tree)\n",
        "dec_tree_accuracy"
      ],
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9963335224755385, 0.5757632598474859]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "NJpd7SKCtjTE"
      },
      "source": [
        "###Linear SVC"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "RP5UcxE5GaI8"
      },
      "source": [
        "The objective of a Linear SVC (Support Vector Classifier) is to fit to the data provided, returning a \"best fit\" hyperplane that divides, or categorizes, data. After getting the hyperplane,the features are fed to the classifier to see what the \"predicted\" class is. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "kOQ0SQvmqtRU",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "b693713b-16f2-4ab1-92b2-29d594e82cc7"
      },
      "source": [
        "linear_sv = LinearSVC()\n",
        "\n",
        "linear_sv_accuracy = accuracy(linear_sv)\n",
        "linear_sv_accuracy"
      ],
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9917346196668266, 0.7177448197054038]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 57
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "Y7GmjutXf0De"
      },
      "source": [
        "##Plot all models and respective accuracy scores"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "Mul75ld8gFBy"
      },
      "source": [
        "Below is a graph of all models which allows for simplified comparison\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "7g0ATkIhq1-I",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 294
        },
        "outputId": "1af726fa-cf1e-49f8-9b44-e370a31916a6"
      },
      "source": [
        "models = ['SVC', 'Bernoulli', 'Multinomial Naive Bayes', 'SGDClassifier', 'Random Forest', 'KNearestNeighbours', 'Gradient Booster', 'Extra Trees', 'Bagging', 'Decision Tree', 'Linear SV']\n",
        "bar_widths = [sv_accuracy[1], bernoulli_accuracy[1], mnb_accuracy[1], sgd_accuracy[1], rand_forest_accuracy[1], knn_accuracy[1], grad_booster_accuracy[1], extra_trees_accuracy[1], bagging_accuracy[1], dec_tree_accuracy[1], linear_sv_accuracy[1]]\n",
        "bar_positions = arange(11) + 0.75\n",
        "tick_positions = range(1,12)\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "ax.barh(bar_positions, bar_widths, 0.5)\n",
        "ax.set_yticks(tick_positions)\n",
        "ax.set_yticklabels(models)\n",
        "\n",
        "ax.set_ylabel('Model')\n",
        "ax.set_xlabel('Accuracy')\n",
        "ax.set_title('Accuracy For Each Model Trained')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "z9AGIAIFl7o4"
      },
      "source": [
        "###Fit best models and use the metrics module"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Eco_i27gmE25",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        },
        "outputId": "60c4c3ec-a33b-41b6-87f3-51c62b95ddae"
      },
      "source": [
        "linear_sv.fit(X_train, y_train)"
      ],
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
              "          intercept_scaling=1, loss='squared_hinge', max_iter=1000,\n",
              "          multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,\n",
              "          verbose=0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "V_qeRzDdmaIM",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 398
        },
        "outputId": "d47ad5cf-0559-4179-8259-381249305e87"
      },
      "source": [
        "#confusion matrix and classification_report\n",
        "y_pred = linear_sv.predict(X_test)\n",
        "\n",
        "print(confusion_matrix(y_test,y_pred))\n",
        "\n",
        "print('\\n\\nAccuracy score: ' + str(accuracy_score(y_test, y_pred)))\n",
        "print(\"\\n\\nClassification Report:\\n\\n\", classification_report(y_test,y_pred,target_names=['Anti', 'Neutral','Pro','News']))"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[  94   26  145   13]\n",
            " [  11  145  237   32]\n",
            " [  12   47 1559  137]\n",
            " [   3   10  205  488]]\n",
            "\n",
            "\n",
            "Accuracy score: 0.7225031605562579\n",
            "\n",
            "\n",
            "Classification Report:\n",
            "\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "        Anti       0.78      0.34      0.47       278\n",
            "     Neutral       0.64      0.34      0.44       425\n",
            "         Pro       0.73      0.89      0.80      1755\n",
            "        News       0.73      0.69      0.71       706\n",
            "\n",
            "    accuracy                           0.72      3164\n",
            "   macro avg       0.72      0.56      0.61      3164\n",
            "weighted avg       0.72      0.72      0.70      3164\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bd_rTKWuyUNc",
        "colab_type": "text"
      },
      "source": [
        "Below is a visual of the confusion matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Pk_XmVxUj31M",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "sentiment_code = {1:'Pro', 2:'News', 0:'Neutral', -1:'Anti'}"
      ],
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZFtAT6rmj31P",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train['sentiment_code'] = train['sentiment'].map(sentiment_code)"
      ],
      "execution_count": 62,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fhieIQs_j31V",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 404
        },
        "outputId": "a6bfb63b-39b5-459e-823f-c17b81661cf7"
      },
      "source": [
        "aux_train = train[['sentiment', 'sentiment_code']].drop_duplicates().sort_values('sentiment_code')\n",
        "conf_matrix = confusion_matrix(y_test, y_pred)\n",
        "plt.figure(figsize=(12.8,6))\n",
        "sns.heatmap(conf_matrix, \n",
        "            annot=True,\n",
        "            cbar=False,\n",
        "            fmt='g',\n",
        "            xticklabels=aux_train['sentiment'].values, \n",
        "            yticklabels=aux_train['sentiment'].values,\n",
        "            cmap=\"Blues\")\n",
        "plt.ylabel('Predicted')\n",
        "plt.xlabel('Actual')\n",
        "plt.title('Confusion matrix')\n",
        "plt.show()"
      ],
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 921.6x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "rNoRppsOm-Tl",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "953f3343-9e54-426a-bbfe-a4b9f9d5e14d"
      },
      "source": [
        "sgd.fit(X_train,y_train)"
      ],
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "SGDClassifier(alpha=0.0001, average=False, class_weight=None,\n",
              "              early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n",
              "              l1_ratio=0.15, learning_rate='optimal', loss='hinge',\n",
              "              max_iter=1000, n_iter_no_change=5, n_jobs=None, penalty='l2',\n",
              "              power_t=0.5, random_state=None, shuffle=True, tol=0.001,\n",
              "              validation_fraction=0.1, verbose=0, warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 64
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "x0g3bcu3nFNB",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 398
        },
        "outputId": "e880b8a7-dcb3-4379-a1b7-6afdde6ada20"
      },
      "source": [
        "y_predict = sgd.predict(X_test)\n",
        "\n",
        "print(confusion_matrix(y_test,y_predict))\n",
        "\n",
        "print('\\n\\nAccuracy score: ' + str(accuracy_score(y_test, y_pred)))\n",
        "print(\"\\n\\nClassification Report:\\n\\n\", classification_report(y_test,y_predict,target_names=['Anti', 'Neutral','Pro','News']))\n",
        "\n"
      ],
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[  85   18  160   15]\n",
            " [   9  115  269   32]\n",
            " [   7   28 1596  124]\n",
            " [   2    5  222  477]]\n",
            "\n",
            "\n",
            "Accuracy score: 0.7225031605562579\n",
            "\n",
            "\n",
            "Classification Report:\n",
            "\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "        Anti       0.83      0.31      0.45       278\n",
            "     Neutral       0.69      0.27      0.39       425\n",
            "         Pro       0.71      0.91      0.80      1755\n",
            "        News       0.74      0.68      0.70       706\n",
            "\n",
            "    accuracy                           0.72      3164\n",
            "   macro avg       0.74      0.54      0.58      3164\n",
            "weighted avg       0.72      0.72      0.69      3164\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TQVHaJhej31g",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "sentiment_code = {1:'Pro', 2:'News', 0:'Neutral', -1:'Anti'}"
      ],
      "execution_count": 66,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L0hxKMmgj31i",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train['sentiment_code'] = train['sentiment'].map(sentiment_code)"
      ],
      "execution_count": 67,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H-G7iYwFj31n",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 404
        },
        "outputId": "d3d716ca-7938-42e6-e087-e6d3b0198e5a"
      },
      "source": [
        "aux_train = train[['sentiment', 'sentiment_code']].drop_duplicates().sort_values('sentiment_code')\n",
        "conf_matrix = confusion_matrix(y_test, y_predict)\n",
        "plt.figure(figsize=(12.8,6))\n",
        "sns.heatmap(conf_matrix, \n",
        "            annot=True,\n",
        "            cbar=False,\n",
        "            fmt='g',\n",
        "            xticklabels=aux_train['sentiment'].values, \n",
        "            yticklabels=aux_train['sentiment'].values,\n",
        "            cmap=\"Blues\")\n",
        "plt.ylabel('Predicted')\n",
        "plt.xlabel('Actual')\n",
        "plt.title('Confusion matrix')\n",
        "plt.show()"
      ],
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 921.6x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "jjYtUYskQIuf"
      },
      "source": [
        "##Modelling with resampled data\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "u9fh-JG1cmdE"
      },
      "source": [
        "Since there has already been an observation that the data is imbalanced, one must attempt improving the accuracy of the above modelling. \n",
        "\n",
        "\n",
        "This will be done by introducing resampling techniques, Downsampling and SMOTE."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "uO9pmo3oqsA7"
      },
      "source": [
        "###Downsampling"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "i0kYHEYAeAk7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "14bcf0a3-089c-4697-9c23-29ebdd7668cf"
      },
      "source": [
        "train = pd.read_csv('https://raw.githubusercontent.com/Stephane-Masamba/Team_4_CPT_ML-Classification/Mikael/train%20(1).csv')\n",
        "print(train.head())"
      ],
      "execution_count": 69,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "   sentiment                                            message  tweetid\n",
            "0          1  PolySciMajor EPA chief doesn't think carbon di...   625221\n",
            "1          1  It's not like we lack evidence of anthropogeni...   126103\n",
            "2          2  RT @RawStory: Researchers say we have three ye...   698562\n",
            "3          1  #TodayinMaker# WIRED : 2016 was a pivotal year...   573736\n",
            "4          1  RT @SoyNovioDeTodas: It's 2016, and a racist, ...   466954\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "QSCGpar4esjq",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "ed1e9fee-41bd-4b3b-8d2a-6ff4ae093f57"
      },
      "source": [
        "test = pd.read_csv('https://raw.githubusercontent.com/Stephane-Masamba/Team_4_CPT_ML-Classification/Mikael/test%20(1).csv')\n",
        "print(test.head())"
      ],
      "execution_count": 70,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                                             message  tweetid\n",
            "0  Europe will now be looking to China to make su...   169760\n",
            "1  Combine this with the polling of staffers re c...    35326\n",
            "2  The scary, unimpeachable evidence that climate...   224985\n",
            "3  @Karoli @morgfair @OsborneInk @dailykos \\nPuti...   476263\n",
            "4  RT @FakeWillMoore: 'Female orgasms cause globa...   872928\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "_IsTEaNLG0zr"
      },
      "source": [
        "Clean the data by calling the cleaning functions"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "_IZ_5Gv-G-VW",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "7f8a0984-dc10-482b-e8dd-b74ca0d2eee0"
      },
      "source": [
        "clean_text(train)\n",
        "remove_stopwords(train)\n",
        "entities(train)\n",
        "lem_text(train)\n",
        "train"
      ],
      "execution_count": 71,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>sentiment</th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>ORG chief donot think carbon dioxide main caus...</td>\n",
              "      <td>625221</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>like lack evidence anthropogenic global warming</td>\n",
              "      <td>126103</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>rt researcher say DATE act climate change late</td>\n",
              "      <td>698562</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1</td>\n",
              "      <td>todayinmaker wire DATE pivotal year war climat...</td>\n",
              "      <td>573736</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1</td>\n",
              "      <td>rt DATE racist sexist climate change deny bigo...</td>\n",
              "      <td>466954</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15814</th>\n",
              "      <td>1</td>\n",
              "      <td>rt take material global warming lgbt right hea...</td>\n",
              "      <td>22001</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15815</th>\n",
              "      <td>2</td>\n",
              "      <td>rt climate change could break CARDINAL relatio...</td>\n",
              "      <td>17856</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15816</th>\n",
              "      <td>0</td>\n",
              "      <td>person rt nytimesworld trump actually believe ...</td>\n",
              "      <td>384248</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15817</th>\n",
              "      <td>-1</td>\n",
              "      <td>rt hey liberal climate change crap hoax ties a...</td>\n",
              "      <td>819732</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15818</th>\n",
              "      <td>0</td>\n",
              "      <td>rt climate change equation CARDINAL screenshot</td>\n",
              "      <td>806319</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>15819 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       sentiment                                            message  tweetid\n",
              "0              1  ORG chief donot think carbon dioxide main caus...   625221\n",
              "1              1    like lack evidence anthropogenic global warming   126103\n",
              "2              2     rt researcher say DATE act climate change late   698562\n",
              "3              1  todayinmaker wire DATE pivotal year war climat...   573736\n",
              "4              1  rt DATE racist sexist climate change deny bigo...   466954\n",
              "...          ...                                                ...      ...\n",
              "15814          1  rt take material global warming lgbt right hea...    22001\n",
              "15815          2  rt climate change could break CARDINAL relatio...    17856\n",
              "15816          0  person rt nytimesworld trump actually believe ...   384248\n",
              "15817         -1  rt hey liberal climate change crap hoax ties a...   819732\n",
              "15818          0     rt climate change equation CARDINAL screenshot   806319\n",
              "\n",
              "[15819 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 71
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "8WkFVIIEOVJj",
        "colab": {}
      },
      "source": [
        "train_majority = train[train.sentiment== 1]\n",
        "train_0 = train[train.sentiment== 0]\n",
        "train_2 = train[train.sentiment== 2]\n",
        "\n",
        "train_minority = train[train.sentiment==-1]\n",
        "\n"
      ],
      "execution_count": 72,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "LtbPdohfRSh1",
        "colab": {}
      },
      "source": [
        "# Downsample majority classes\n",
        "train_majority_downsampled = resample(train_majority, \n",
        "                                 replace=False,    # sample without replacement\n",
        "                                 n_samples=1296,     # to match minority class\n",
        "                                 random_state=123) # reproducible results\n",
        "\n",
        "\n",
        "train_0_downsampled = resample(train_0, \n",
        "                                 replace=False,    \n",
        "                                 n_samples=1296,     \n",
        "                                 random_state=123) \n",
        "\n",
        "train_2_downsampled = resample(train_2, \n",
        "                                 replace=False,    \n",
        "                                 n_samples=1296,     \n",
        "                                 random_state=123) \n",
        "\n",
        "                      \n",
        "\n",
        "\n",
        "\n",
        "# Combine minority class with downsampled majority class\n",
        "train_downsampled1 = pd.concat([train_0_downsampled,train_2_downsampled])\n",
        "\n",
        "train_downsampled2 = pd.concat([train_majority_downsampled, train_minority])\n",
        "\n",
        "train_downsampled =  pd.concat([train_downsampled1, train_downsampled2])"
      ],
      "execution_count": 73,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Mj9doThtRnqb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "ce978b3c-89b0-4c7c-b5e8-8cb28a3bfcc0"
      },
      "source": [
        "train_downsampled"
      ],
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>sentiment</th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>13470</th>\n",
              "      <td>0</td>\n",
              "      <td>rt really god global warming exist movie pulp ...</td>\n",
              "      <td>574574</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8148</th>\n",
              "      <td>0</td>\n",
              "      <td>also storm global warming maybe probably</td>\n",
              "      <td>761606</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11544</th>\n",
              "      <td>0</td>\n",
              "      <td>rt savor ur cold CARDINALs u climate change OR...</td>\n",
              "      <td>639609</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7282</th>\n",
              "      <td>0</td>\n",
              "      <td>rt -PRON-have begin m-PRON-ss day endless pape...</td>\n",
              "      <td>121132</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12634</th>\n",
              "      <td>0</td>\n",
              "      <td>omg celebritie moral action climate change sui...</td>\n",
              "      <td>744560</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15704</th>\n",
              "      <td>-1</td>\n",
              "      <td>rt point wthe delusional leave literally every...</td>\n",
              "      <td>890582</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15705</th>\n",
              "      <td>-1</td>\n",
              "      <td>person admit GPE accord willnot solve issue cl...</td>\n",
              "      <td>380307</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15711</th>\n",
              "      <td>-1</td>\n",
              "      <td>rt funny people believe climate change bc scie...</td>\n",
              "      <td>390339</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15806</th>\n",
              "      <td>-1</td>\n",
              "      <td>rt PERSON idiot think climate change cause pro...</td>\n",
              "      <td>213863</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15817</th>\n",
              "      <td>-1</td>\n",
              "      <td>rt hey liberal climate change crap hoax ties a...</td>\n",
              "      <td>819732</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5184 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       sentiment                                            message  tweetid\n",
              "13470          0  rt really god global warming exist movie pulp ...   574574\n",
              "8148           0           also storm global warming maybe probably   761606\n",
              "11544          0  rt savor ur cold CARDINALs u climate change OR...   639609\n",
              "7282           0  rt -PRON-have begin m-PRON-ss day endless pape...   121132\n",
              "12634          0  omg celebritie moral action climate change sui...   744560\n",
              "...          ...                                                ...      ...\n",
              "15704         -1  rt point wthe delusional leave literally every...   890582\n",
              "15705         -1  person admit GPE accord willnot solve issue cl...   380307\n",
              "15711         -1  rt funny people believe climate change bc scie...   390339\n",
              "15806         -1  rt PERSON idiot think climate change cause pro...   213863\n",
              "15817         -1  rt hey liberal climate change crap hoax ties a...   819732\n",
              "\n",
              "[5184 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 74
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "JOZcGngcRrgF",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        },
        "outputId": "f92937ad-db76-4600-fe15-6722c01cebd6"
      },
      "source": [
        "train_downsampled['sentiment'].value_counts()"
      ],
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "-1    1296\n",
              " 2    1296\n",
              " 1    1296\n",
              " 0    1296\n",
              "Name: sentiment, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 75
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "yltvCR60E-gT",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 278
        },
        "outputId": "fb87c2dc-b39e-4eb4-ed7a-9806b5f72818"
      },
      "source": [
        "counts = train[\"sentiment\"].value_counts()\n",
        "counti = train_downsampled['sentiment'].value_counts()\n",
        "\n",
        "plt.bar(range(len(counts)), counts)\n",
        "plt.bar(range(len(counts)),counti,color='red')\n",
        "plt.xticks([0, 1, 2, 3], ['Pro', 'News', 'Neutral', 'Anti'])\n",
        "\n",
        "plt.ylabel(\"Total per class\")\n",
        "plt.xlabel(\"Sentiment Classes\")\n",
        "plt.legend(['original','resampled'])\n",
        "plt.show()"
      ],
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "sP8F20Iyj23V"
      },
      "source": [
        "Now that the data is balanced, we go ahead and build a model. This time we check the performance with onlyy the two best performing models as per the above graph."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "MJFyB85JgBHH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "outputId": "e8d9632b-436b-43e1-9377-7764c515c87e"
      },
      "source": [
        "X_down = train['message']\n",
        "X_down"
      ],
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0        ORG chief donot think carbon dioxide main caus...\n",
              "1          like lack evidence anthropogenic global warming\n",
              "2           rt researcher say DATE act climate change late\n",
              "3        todayinmaker wire DATE pivotal year war climat...\n",
              "4        rt DATE racist sexist climate change deny bigo...\n",
              "                               ...                        \n",
              "15814    rt take material global warming lgbt right hea...\n",
              "15815    rt climate change could break CARDINAL relatio...\n",
              "15816    person rt nytimesworld trump actually believe ...\n",
              "15817    rt hey liberal climate change crap hoax ties a...\n",
              "15818       rt climate change equation CARDINAL screenshot\n",
              "Name: message, Length: 15819, dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 77
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "P2wsrp-YgQip",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "outputId": "d36d7969-4d3a-4505-e4f4-a2486407ef15"
      },
      "source": [
        "y_down = train['sentiment']\n",
        "y_down"
      ],
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0        1\n",
              "1        1\n",
              "2        2\n",
              "3        1\n",
              "4        1\n",
              "        ..\n",
              "15814    1\n",
              "15815    2\n",
              "15816    0\n",
              "15817   -1\n",
              "15818    0\n",
              "Name: sentiment, Length: 15819, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 78
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "TRLvUU0qu5Hb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "39a00427-4e88-4d69-a57d-2ca3bb745590"
      },
      "source": [
        "X_down = tf_vecto.fit_transform(X_down)\n",
        "X_down"
      ],
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<15819x74310 sparse matrix of type '<class 'numpy.float64'>'\n",
              "\twith 288784 stored elements in Compressed Sparse Row format>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 79
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "dn-C8qT9hA6n",
        "colab": {}
      },
      "source": [
        "#train_test_split\n",
        "X_train1,X_test1,y_train1,y_test1 = train_test_split(X_down,y_down,test_size=0.2,random_state=0)"
      ],
      "execution_count": 80,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "yMS2y9rGulE5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        },
        "outputId": "a4dc85ed-d9bc-4176-e582-a24e2459f2ad"
      },
      "source": [
        "\n",
        "lsvm = LinearSVC()\n",
        "lsvm.fit(X_train1, y_train1)"
      ],
      "execution_count": 81,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
              "          intercept_scaling=1, loss='squared_hinge', max_iter=1000,\n",
              "          multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,\n",
              "          verbose=0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 81
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "TXYesYXKvMM2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 398
        },
        "outputId": "1a231862-cdff-4219-c38d-ce53cfe72d67"
      },
      "source": [
        "#confusion matrix and classification_report\n",
        "y_pred1 = lsvm.predict(X_test1)\n",
        "\n",
        "print(confusion_matrix(y_test1,y_pred1))\n",
        "\n",
        "print('\\n\\nAccuracy score: ' + str(accuracy_score(y_test1, y_pred1)))\n",
        "print(\"\\n\\nClassification Report:\\n\\n\",classification_report(y_test1,y_pred1,target_names=['Anti', 'Neutral','Pro','News']))"
      ],
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[  84   23  128   14]\n",
            " [  10  146  265   52]\n",
            " [  12   47 1514  128]\n",
            " [   3   13  197  528]]\n",
            "\n",
            "\n",
            "Accuracy score: 0.718078381795196\n",
            "\n",
            "\n",
            "Classification Report:\n",
            "\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "        Anti       0.77      0.34      0.47       249\n",
            "     Neutral       0.64      0.31      0.42       473\n",
            "         Pro       0.72      0.89      0.80      1701\n",
            "        News       0.73      0.71      0.72       741\n",
            "\n",
            "    accuracy                           0.72      3164\n",
            "   macro avg       0.71      0.56      0.60      3164\n",
            "weighted avg       0.71      0.72      0.70      3164\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dN8ur5dSzaFO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "sentiment_code = {1:'Pro', 2:'News', 0:'Neutral', -1:'Anti'}"
      ],
      "execution_count": 83,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5CLzTodGzmMs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train['sentiment_code'] = train['sentiment'].map(sentiment_code)"
      ],
      "execution_count": 84,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PC3NblX7zKMY",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 404
        },
        "outputId": "c24bb6bf-f2e6-403c-e21d-5355a6150fa1"
      },
      "source": [
        "aux_train = train[['sentiment', 'sentiment_code']].drop_duplicates().sort_values('sentiment_code')\n",
        "conf_matrix = confusion_matrix(y_test1, y_pred1)\n",
        "plt.figure(figsize=(12.8,6))\n",
        "sns.heatmap(conf_matrix, \n",
        "            annot=True,\n",
        "            cbar=False,\n",
        "            fmt='g',\n",
        "            xticklabels=aux_train['sentiment'].values, \n",
        "            yticklabels=aux_train['sentiment'].values,\n",
        "            cmap=\"Blues\")\n",
        "plt.ylabel('Predicted')\n",
        "plt.xlabel('Actual')\n",
        "plt.title('Confusion matrix')\n",
        "plt.show()"
      ],
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 921.6x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "1PXmOs5S5Ycq"
      },
      "source": [
        "###SMOTE "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "380q1LzJh0Oc"
      },
      "source": [
        "Using data from the first data split"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "tvBjg1eUYe5T",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "ccd1b710-404e-4f96-87f2-aa0a8bf270b6"
      },
      "source": [
        "print(X_train.shape,y_train.shape)"
      ],
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(12655, 74310) (12655,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "D_431FaxZikW",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "43652268-0ce8-4bf9-aa39-8be1d0bb0a12"
      },
      "source": [
        "smote = SMOTE(\"minority\")\n",
        "X_sm , y_sm = smote.fit_resample(X_train,y_train)"
      ],
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/utils/deprecation.py:87: FutureWarning: Function safe_indexing is deprecated; safe_indexing is deprecated in version 0.22 and will be removed in version 0.24.\n",
            "  warnings.warn(msg, category=FutureWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "G7B3DarxZoG1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "3c07b54b-1920-4c95-b13c-aafb8ae85465"
      },
      "source": [
        "print(X_sm.shape,y_sm.shape)"
      ],
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(18412, 74310) (18412,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "LKKuNpMzZyBW",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 398
        },
        "outputId": "4f3dcfc4-5cc4-4fc8-a9bc-e2f5c9972315"
      },
      "source": [
        "ls= LinearSVC()\n",
        "ls.fit(X_sm, y_sm)\n",
        "\n",
        "\n",
        "#confusion matrix and classification_report\n",
        "y_predsm = ls.predict(X_test)\n",
        "\n",
        "print(confusion_matrix(y_test,y_predsm))\n",
        "\n",
        "print('\\n\\nAccuracy score: ' + str(accuracy_score(y_test, y_predsm)))\n",
        "print(\"\\n\\nClassification Report:\\n\\n\",classification_report(y_test,y_predsm,target_names=['Anti', 'Neutral','Pro','News']))"
      ],
      "execution_count": 89,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[ 122   18  127   11]\n",
            " [  37  120  238   30]\n",
            " [  41   46 1535  133]\n",
            " [  11   12  201  482]]\n",
            "\n",
            "\n",
            "Accuracy score: 0.7139696586599241\n",
            "\n",
            "\n",
            "Classification Report:\n",
            "\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "        Anti       0.58      0.44      0.50       278\n",
            "     Neutral       0.61      0.28      0.39       425\n",
            "         Pro       0.73      0.87      0.80      1755\n",
            "        News       0.73      0.68      0.71       706\n",
            "\n",
            "    accuracy                           0.71      3164\n",
            "   macro avg       0.66      0.57      0.60      3164\n",
            "weighted avg       0.70      0.71      0.70      3164\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R5IWO69qj32o",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "sentiment_code = {1:'Pro', 2:'News', 0:'Neutral', -1:'Anti'}"
      ],
      "execution_count": 90,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hPBvfTdkj32r",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train['sentiment_code'] = train['sentiment'].map(sentiment_code)"
      ],
      "execution_count": 91,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "95FFi172j320",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 404
        },
        "outputId": "a0dd75cc-7bb3-4d09-abbf-ff0fdbd04a43"
      },
      "source": [
        "aux_train = train[['sentiment', 'sentiment_code']].drop_duplicates().sort_values('sentiment_code')\n",
        "conf_matrix = confusion_matrix(y_test, y_predsm)\n",
        "plt.figure(figsize=(12.8,6))\n",
        "sns.heatmap(conf_matrix, \n",
        "            annot=True,\n",
        "            cbar=False,\n",
        "            fmt='g',\n",
        "            xticklabels=aux_train['sentiment'].values, \n",
        "            yticklabels=aux_train['sentiment'].values,\n",
        "            cmap=\"Blues\")\n",
        "plt.ylabel('Predicted')\n",
        "plt.xlabel('Actual')\n",
        "plt.title('Confusion matrix')\n",
        "plt.show()"
      ],
      "execution_count": 92,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 921.6x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "IM-yC7gnkay5"
      },
      "source": [
        "##Modelling with minimal data cleaning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "rcAxGtQ9iSId"
      },
      "source": [
        "To conclude the modelling section, we will explore what one could have done by intuition. This is just to check how accurate our model would be if we did not introduce the 'fancy' cleaning functions."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Brtlh1UttRzv",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "25e2be00-2b46-41e8-ad7f-6ddf633f0eac"
      },
      "source": [
        "train = pd.read_csv('https://raw.githubusercontent.com/Stephane-Masamba/Team_4_CPT_ML-Classification/Mikael/train%20(1).csv')\n",
        "print(train.head())"
      ],
      "execution_count": 93,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "   sentiment                                            message  tweetid\n",
            "0          1  PolySciMajor EPA chief doesn't think carbon di...   625221\n",
            "1          1  It's not like we lack evidence of anthropogeni...   126103\n",
            "2          2  RT @RawStory: Researchers say we have three ye...   698562\n",
            "3          1  #TodayinMaker# WIRED : 2016 was a pivotal year...   573736\n",
            "4          1  RT @SoyNovioDeTodas: It's 2016, and a racist, ...   466954\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "vl3SL2fotbpK",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "4d219b41-b6f4-45ae-d914-60d8e23d8220"
      },
      "source": [
        "test = pd.read_csv('https://raw.githubusercontent.com/Stephane-Masamba/Team_4_CPT_ML-Classification/Mikael/test%20(1).csv')\n",
        "print(test.head())\n"
      ],
      "execution_count": 94,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "                                             message  tweetid\n",
            "0  Europe will now be looking to China to make su...   169760\n",
            "1  Combine this with the polling of staffers re c...    35326\n",
            "2  The scary, unimpeachable evidence that climate...   224985\n",
            "3  @Karoli @morgfair @OsborneInk @dailykos \\nPuti...   476263\n",
            "4  RT @FakeWillMoore: 'Female orgasms cause globa...   872928\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "kHFkpesDthSE",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 121
        },
        "outputId": "118bbc2a-e134-4d26-9e68-d91225b882ea"
      },
      "source": [
        "sample = pd.read_csv('https://raw.githubusercontent.com/Stephane-Masamba/Team_4_CPT_ML-Classification/Mikael/sample_submission.csv')\n",
        "print(sample.head())\n"
      ],
      "execution_count": 95,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "   tweetid  sentiment\n",
            "0   169760          1\n",
            "1    35326          1\n",
            "2   224985          1\n",
            "3   476263          1\n",
            "4   872928          1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "95gOu5vo2VVV"
      },
      "source": [
        "This time we only call one cleaning function which is the clean_text() function. This does not include tokenization as well as lemmatization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Mqw1jzqUDL3p",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "7a06ad47-6cf8-4b07-ba87-cb01d8dfc935"
      },
      "source": [
        "clean_text(train)\n",
        "remove_stopwords(train)\n",
        "train"
      ],
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>sentiment</th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>polyscimajor epa chief doesnt think carbon dio...</td>\n",
              "      <td>625221</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>like lack evidence anthropogenic global warming</td>\n",
              "      <td>126103</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>rt researchers say three years act climate cha...</td>\n",
              "      <td>698562</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1</td>\n",
              "      <td>todayinmaker wired 2016 pivotal year war clima...</td>\n",
              "      <td>573736</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1</td>\n",
              "      <td>rt 2016 racist sexist climate change denying b...</td>\n",
              "      <td>466954</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15814</th>\n",
              "      <td>1</td>\n",
              "      <td>rt took material global warming lgbt rights he...</td>\n",
              "      <td>22001</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15815</th>\n",
              "      <td>2</td>\n",
              "      <td>rt climate change could breaking 200millionyea...</td>\n",
              "      <td>17856</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15816</th>\n",
              "      <td>0</td>\n",
              "      <td>notiven rt nytimesworld trump actually believe...</td>\n",
              "      <td>384248</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15817</th>\n",
              "      <td>-1</td>\n",
              "      <td>rt hey liberals climate change crap hoax ties ...</td>\n",
              "      <td>819732</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15818</th>\n",
              "      <td>0</td>\n",
              "      <td>rt climate change equation 4 screenshots</td>\n",
              "      <td>806319</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>15819 rows × 3 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "       sentiment                                            message  tweetid\n",
              "0              1  polyscimajor epa chief doesnt think carbon dio...   625221\n",
              "1              1    like lack evidence anthropogenic global warming   126103\n",
              "2              2  rt researchers say three years act climate cha...   698562\n",
              "3              1  todayinmaker wired 2016 pivotal year war clima...   573736\n",
              "4              1  rt 2016 racist sexist climate change denying b...   466954\n",
              "...          ...                                                ...      ...\n",
              "15814          1  rt took material global warming lgbt rights he...    22001\n",
              "15815          2  rt climate change could breaking 200millionyea...    17856\n",
              "15816          0  notiven rt nytimesworld trump actually believe...   384248\n",
              "15817         -1  rt hey liberals climate change crap hoax ties ...   819732\n",
              "15818          0           rt climate change equation 4 screenshots   806319\n",
              "\n",
              "[15819 rows x 3 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 96
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "IoOaL19lEO4C",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "a1877542-7f16-4af4-9a46-d2b5e948438f"
      },
      "source": [
        "clean_text(test)\n",
        "remove_stopwords(test)\n",
        "test"
      ],
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>message</th>\n",
              "      <th>tweetid</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>europe looking china make sure alone fighting ...</td>\n",
              "      <td>169760</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>combine polling staffers climate change womens...</td>\n",
              "      <td>35326</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>scary unimpeachable evidence climate change al...</td>\n",
              "      <td>224985</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>putin got jill trump doesnt believe climate ch...</td>\n",
              "      <td>476263</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>rt female orgasms cause global warming sarcast...</td>\n",
              "      <td>872928</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10541</th>\n",
              "      <td>rt brb writing poem climate change climatechan...</td>\n",
              "      <td>895714</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10542</th>\n",
              "      <td>2016 year climate change came home hottest yea...</td>\n",
              "      <td>875167</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10543</th>\n",
              "      <td>rt pacific countries positive fiji leading glo...</td>\n",
              "      <td>78329</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10544</th>\n",
              "      <td>rt youre hot must cause global warming aldubla...</td>\n",
              "      <td>867455</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10545</th>\n",
              "      <td>rt climate change global issue thats getting w...</td>\n",
              "      <td>470892</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>10546 rows × 2 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "                                                 message  tweetid\n",
              "0      europe looking china make sure alone fighting ...   169760\n",
              "1      combine polling staffers climate change womens...    35326\n",
              "2      scary unimpeachable evidence climate change al...   224985\n",
              "3      putin got jill trump doesnt believe climate ch...   476263\n",
              "4      rt female orgasms cause global warming sarcast...   872928\n",
              "...                                                  ...      ...\n",
              "10541  rt brb writing poem climate change climatechan...   895714\n",
              "10542  2016 year climate change came home hottest yea...   875167\n",
              "10543  rt pacific countries positive fiji leading glo...    78329\n",
              "10544  rt youre hot must cause global warming aldubla...   867455\n",
              "10545  rt climate change global issue thats getting w...   470892\n",
              "\n",
              "[10546 rows x 2 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 97
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "PW0Aqhr2kdeF"
      },
      "source": [
        "###Features and Modelling"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "f19MFMIY3KLi",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "outputId": "035f7d29-75e6-46f3-9585-24725468589a"
      },
      "source": [
        "X_min = train['message']\n",
        "X_min\n"
      ],
      "execution_count": 98,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0        polyscimajor epa chief doesnt think carbon dio...\n",
              "1          like lack evidence anthropogenic global warming\n",
              "2        rt researchers say three years act climate cha...\n",
              "3        todayinmaker wired 2016 pivotal year war clima...\n",
              "4        rt 2016 racist sexist climate change denying b...\n",
              "                               ...                        \n",
              "15814    rt took material global warming lgbt rights he...\n",
              "15815    rt climate change could breaking 200millionyea...\n",
              "15816    notiven rt nytimesworld trump actually believe...\n",
              "15817    rt hey liberals climate change crap hoax ties ...\n",
              "15818             rt climate change equation 4 screenshots\n",
              "Name: message, Length: 15819, dtype: object"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 98
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "K_6P14uc3W0H",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "outputId": "928cf0be-380e-49a7-d05a-06875ad99347"
      },
      "source": [
        "y_min = train['sentiment']\n",
        "y_min\n"
      ],
      "execution_count": 99,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0        1\n",
              "1        1\n",
              "2        2\n",
              "3        1\n",
              "4        1\n",
              "        ..\n",
              "15814    1\n",
              "15815    2\n",
              "15816    0\n",
              "15817   -1\n",
              "15818    0\n",
              "Name: sentiment, Length: 15819, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 99
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "ZpHXHlui3Zws"
      },
      "source": [
        "###Convert to machine language"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "lD2xQ8Fb5sVw",
        "colab": {}
      },
      "source": [
        "X_min = tf_vecto.fit_transform(X_min)\n"
      ],
      "execution_count": 100,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "xkBBEQStmu3i",
        "colab": {}
      },
      "source": [
        "#train_test_split\n",
        "X_train2,X_test2,y_train2,y_test2 = train_test_split(X_min,y_min,test_size=0.2,random_state=0)"
      ],
      "execution_count": 101,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "s8DgqL504f16"
      },
      "source": [
        "###Stochastic Gradient Descent"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "3NzmKTgzIxgh",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "8b2643f9-5319-4eaf-875a-17e1ed4b057d"
      },
      "source": [
        "sgd = SGDClassifier()\n",
        "\n",
        "sgd_accuracy = accuracy(sgd)\n",
        "sgd_accuracy\n",
        "\n"
      ],
      "execution_count": 102,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9517036796816816, 0.7283645085619799]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 102
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "ONjiNYLnmbq7"
      },
      "source": [
        "###Linear SVC"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "OqtzXN2aws54",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "bf7805de-8bf5-45ee-de05-5760a3945487"
      },
      "source": [
        "linear_sv = LinearSVC()\n",
        "\n",
        "linear_sv_accuracy = accuracy(linear_sv)\n",
        "linear_sv_accuracy\n"
      ],
      "execution_count": 103,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.9938365281882657, 0.7289332488120185]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 103
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "EktlZbOIzGsc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        },
        "outputId": "95c7af3a-ba26-40c2-8c2e-8f3768e002f1"
      },
      "source": [
        "linear_sv.fit(X_train2, y_train2)\n"
      ],
      "execution_count": 104,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LinearSVC(C=1.0, class_weight=None, dual=True, fit_intercept=True,\n",
              "          intercept_scaling=1, loss='squared_hinge', max_iter=1000,\n",
              "          multi_class='ovr', penalty='l2', random_state=None, tol=0.0001,\n",
              "          verbose=0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 104
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "dt1k_nN3oFfU",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 398
        },
        "outputId": "763cfa80-8be7-4a13-a06f-de67295aaf7d"
      },
      "source": [
        "#confusion matrix and classification_report\n",
        "y_pred2 = linear_sv.predict(X_test2)\n",
        "\n",
        "\n",
        "print(confusion_matrix(y_test1,y_pred1))\n",
        "\n",
        "print('\\n\\nAccuracy score: ' + str(accuracy_score(y_test2, y_pred1)))\n",
        "print(\"\\n\\nClassification Report:\\n\\n\",classification_report(y_test2,y_pred2,target_names=['Anti', 'Neutral','Pro','News']))"
      ],
      "execution_count": 105,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[  84   23  128   14]\n",
            " [  10  146  265   52]\n",
            " [  12   47 1514  128]\n",
            " [   3   13  197  528]]\n",
            "\n",
            "\n",
            "Accuracy score: 0.718078381795196\n",
            "\n",
            "\n",
            "Classification Report:\n",
            "\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "        Anti       0.73      0.31      0.44       249\n",
            "     Neutral       0.65      0.32      0.42       473\n",
            "         Pro       0.72      0.90      0.80      1701\n",
            "        News       0.76      0.72      0.74       741\n",
            "\n",
            "    accuracy                           0.72      3164\n",
            "   macro avg       0.72      0.56      0.60      3164\n",
            "weighted avg       0.72      0.72      0.70      3164\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FmdjS4gpj33o",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "sentiment_code = {1:'Pro', 2:'News', 0:'Neutral', -1:'Anti'}"
      ],
      "execution_count": 106,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eFgdiDP7j33t",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train['sentiment_code'] = train['sentiment'].map(sentiment_code)"
      ],
      "execution_count": 107,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "92iepoTFj33z",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 404
        },
        "outputId": "76f7dcf8-f26d-4477-e172-7de3f2ce9f91"
      },
      "source": [
        "aux_train = train[['sentiment', 'sentiment_code']].drop_duplicates().sort_values('sentiment_code')\n",
        "conf_matrix = confusion_matrix(y_test1, y_pred1)\n",
        "plt.figure(figsize=(12.8,6))\n",
        "sns.heatmap(conf_matrix, \n",
        "            annot=True,\n",
        "            cbar=False,\n",
        "            fmt='g',\n",
        "            xticklabels=aux_train['sentiment'].values, \n",
        "            yticklabels=aux_train['sentiment'].values,\n",
        "            cmap=\"Blues\")\n",
        "plt.ylabel('Predicted')\n",
        "plt.xlabel('Actual')\n",
        "plt.title('Confusion matrix')\n",
        "plt.show()"
      ],
      "execution_count": 108,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAvcAAAGDCAYAAACvJD/xAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd5hU5d3G8e+PXZQmyNIEbGisUTB20diIGhWVGHuJGrtiwx5RYkk0xt6CiC0WsARQkIhRQQSNisYSu0aNioABEZGl7PK8f+ywL4iwu8BylrPfz3XtxZzznDlzjw5w88wzZyKlhCRJkqQVX4OsA0iSJElaNiz3kiRJUk5Y7iVJkqScsNxLkiRJOWG5lyRJknLCci9JkiTlhOVeklZQEdE4IoZGxLcR8chSnOfwiHhqWWbLSkT8PCLezzqHJGUlvM69JNWuiDgM6AVsCHwHvA78IaU0ZinPeyRwGtA1pVS21EHruIhIwHoppY+yziJJdZUz95JUiyKiF3AD8EegHbAmcBuw3zI4/VrAB/Wh2FdHRBRnnUGSsma5l6RaEhEtgMuAU1NKg1JK36eU5qSUhqaUzi0cs3JE3BAR4ws/N0TEyoWxnSPii4g4OyImRcRXEXFMYexS4BLg4IiYHhHHRsTvI+L++R5/7YhI80pvRBwdEf+JiO8i4pOIOHy+/WPmu1/XiHilsNznlYjoOt/YqIi4PCLGFs7zVES0XsTzn5f/vPny94iIvSLig4iYEhG/m+/4rSPixYiYWjj2lohYqTA2unDYG4Xne/B85z8/IiYAd8/bV7jPuoXH2Lyw3SEivo6InZfqf6wk1WGWe0mqPdsBjYDBiznmImBbYDOgC7A10Hu+8dWAFkBH4Fjg1ohomVLqQ8W7AQ+llJqllO5cXJCIaArcBOyZUloF6ErF8qAfHlcCPFE4thVwHfBERLSa77DDgGOAtsBKwDmLeejVqPhv0JGKf4zcARwBbAH8HLg4IjoVji0HzgJaU/HfrhtwCkBKacfCMV0Kz/eh+c5fQsW7GCfM/8AppY+B84H7I6IJcDdwb0pp1GLyStIKzXIvSbWnFfC/KpbNHA5cllKalFL6GrgUOHK+8TmF8TkppeHAdGCDJcwzF9gkIhqnlL5KKb39I8fsDXyYUrovpVSWUhoAvAfsM98xd6eUPkgplQIPU/EPk0WZQ8XnC+YAA6ko7jemlL4rPP47VPyjhpTSqymlfxYe91PgdmCnajynPimlWYU8C0gp3QF8BLwEtKfiH1OSlFuWe0mqPZOB1lWsBe8AfDbf9meFfZXn+ME/DmYAzWoaJKX0PXAwcBLwVUQ8EREbViPPvEwd59ueUIM8k1NK5YXb88r3xPnGS+fdPyLWj4hhETEhIqZR8c7Ejy75mc/XKaWZVRxzB7AJcHNKaVYVx0rSCs1yL0m150VgFtBjMceMp2JJyTxrFvYtie+BJvNtrzb/YEppREppNypmsN+jovRWlWdepi+XMFNN/IWKXOullJoDvwOiivss9pJvEdGMig803wn8vrDsSJJyy3IvSbUkpfQtFevMby18kLRJRDSMiD0j4urCYQOA3hHRpvDB1EuA+xd1ziq8DuwYEWsWPsx74byBiGgXEfsV1t7PomJ5z9wfOcdwYP2IOCwiiiPiYGBjYNgSZqqJVYBpwPTCuwon/2B8IrBODc95IzAupXQcFZ8l6LvUKSWpDrPcS1ItSildS8U17nsDXwOfAz2BIYVDrgDGAW8CbwGvFfYtyWP9A3iocK5XWbCQNyjkGA9MoWIt+w/LMymlyUB34GwqlhWdB3RPKf1vSTLV0DlUfFj3OyreVXjoB+O/B+4tXE3noKpOFhH7Ab/k/59nL2DzeVcJkqQ88kusJEmSpJxw5l6SJEnKCcu9JEmSlBOWe0mSJCknLPeSJElSTljuJUmSpJxY3LcmZmrK9+VexkcrvEYN/fez8mFaaVnVB0l1XIsmDbOOIC0TjRsu+gv+bB6SJElSTljuJUmSpJyw3EuSJEk5YbmXJEmScsJyL0mSJOWE5V6SJEnKCcu9JEmSlBOWe0mSJCknLPeSJElSTljuJUmSpJyw3EuSJEk5YbmXJEmScsJyL0mSJOWE5V6SJEnKCcu9JEmSlBOWe0mSJCknLPeSJElSTljuJUmSpJyw3EuSJEk5YbmXJEmScsJyL0mSJOWE5V6SJEnKCcu9JEmSlBOWe0mSJCknLPeSJElSTljuJUmSpJyw3EuSJEk5YbmXJEmScsJyL0mSJOWE5V6SJEnKCcu9JEmSlBOWe0mSJCknLPeSJElSTljuJUmSpJyw3EuSJEk5UZx1AC0/A+6/l6FDHiUiWPcn63PR7//AyiuvDMB1V/+BYY8N4tmxr2acUlq0CRO+4uLfnc/kyZOJCH59wEEcdsRvuPXmG3lu5DNEgwaUlJRw6RVX0rZtu6zjSgu46vLevDhmNC1blnDPwCEA/OWma3jh+ecoblhMh45rcMElV7DKKs0pK5vD1Vf04YP336W8vIw99tqXI44+PuNnIC2sT+8LGT16FCUlrfjbkGELjP31nru47po/MfL5F2nZsiSjhPWPM/f1xKRJE3lk4P3cdf8jPPDI45TPLefpEcMBePedf/PdtGkZJ5SqVlRURK9zzmfQY0/w1wcG8tDAB/j444846phjeXjQ4zz06BB+vtPO9Ot7W9ZRpYXsuXcP/nxj3wX2bbn1dtw9YDB3PziYNdZcmwfu6Q/AyKefYs6c2dwzYDB3/PVhhg5+hK/Gf5lFbGmx9u2xP7f17b/Q/glffcWLL4ylffsOGaSq3yz39Uh5eTmzZs2krKyMmaUzad2mLeXl5dxywzWcesY5WceTqtSmTVs22vinADRt2oxOndbl64kTadasWeUxpaWlRERWEaVF6rL5lqzSvMUC+7badnuKiyveRN94k858PWkiABFBaWkpZWVlzJo5i+LihjRt2myhc0pZ22LLrWjeosVC+6+5+krO7HUu+Ofxcrfcl+VERLOU0vTl/bj1Xdu27TjsyGP41V7dWHnlRmy9XVe22W57HnrwPnbYcRdat2mTdUSpRsZ/+QXvv/cum3TuAsAtN13PsMcfo9kqq9DvznszTifV3PChg9l1t18CsHO33Rg7+ln232sXZs2cyalnnfejBUqqi0Y++zRt2rZlgw03zDpKvZTFzP07ixqIiBMiYlxEjLv3rjuWZ6bcmzbtW54f9Sx/G/YPho4YxczSUoYPe4xnnx7BgYccnnU8qUZmzPiec846nXPOv7By1r7n6Wfx5NOj2HPv7jw04P6ME0o1c99dt1NUVMRuv+wOwLtvv0WDBkUMGv4sA4c8ycMP3Mv4Lz/POKVUtdLSUu6843ZO6XlG1lHqrVqZuY+IXosaAhb5vmJKqR/QD2DK9+WpFqLVW6+89CLtO3as/EDLTrvuRv++tzBr1kwO3K9ipmjmzJkcsO8ePPr4iCyjSos1Z84czjnrdPbcex+6/WL3hcb32nsfTjvlRE4+9fQM0kk19/dhQ3hhzGiuv61/5ZKyp0cMZ+vttqe4uCEtS1qxSZfNeO+dt+nQcY2M00qL98Xn/+XLL7/goF/vB8CkiRM49MD9uX/gI7Ru7SqB5aG2luX8EfgzUPYjY67zz8Bqq7Xn7bfeYGZpKSs3asS4l//JoUccxYGHHFF5zK7bb2GxV52WUuLSPr3ptM66HHnUMZX7P/vsU9Zaa20ARj37DGt36pRRQqlmXnpxDAPuu4ub+t5Do0aNK/e3a9ee18a9zB577Utp6Qze+febHHjIkRkmlapnvfU3YOToFyu399x9Vx586FGvlrMc1Va5fw0YklJa6LqKEXFcLT2mFuOnm3Zhl267c9ThB1BcVMT6G2zEfvsflHUsqUZe/9drPDH0MdZbb30OPqAHULEcZ8jgR/ns009pEEH7Dh246OJLM04qLezS3ufy+quv8O3UqRzQvRvHHH8KD9zbn9mzZ3N2z4rLXG68SWfOvrAPPQ48lKsu681RB+9HIrFn9x6su94GGT8DaWEXnNuLca+8zNSp37B7tx05+ZTT+NWvD8w6Vr0WKS371S8RsQEwOaX0v/n2rZZSmhAR7VJKE6s6h8tylAeNGvpGlfJhWumPvRErrVhaNGmYdQRpmWjckEVehqhWmkdK6f35i33B8MJYlcVekiRJUs0tz2lFL3QqSZIk1aLlWe69tqUkSZJUi5ZbuU8p+X3wkiRJUi3y036SJElSTljuJUmSpJyw3EuSJEk5YbmXJEmScsJyL0mSJOWE5V6SJEnKCcu9JEmSlBOWe0mSJCknLPeSJElSTljuJUmSpJyw3EuSJEk5YbmXJEmScsJyL0mSJOWE5V6SJEnKCcu9JEmSlBOWe0mSJCknLPeSJElSTljuJUmSpJyw3EuSJEk5YbmXJEmScsJyL0mSJOWE5V6SJEnKCcu9JEmSlBOWe0mSJCknLPeSJElSTljuJUmSpJyw3EuSJEk5YbmXJEmScsJyL0mSJOWE5V6SJEnKCcu9JEmSlBOWe0mSJCknirMOsCiNVyrKOoK01KZ8PzvrCNIy8e2MOVlHkJZak5XtFsqHxg0XPT/vzL0kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknCjOOoCWjz69L2T06FGUlLTib0OGAfDtt1M57+yzGD/+Szp06Mifr72B5i1aZJxUWtjVl1/MP8eOZtWWJdw1YPACYw8/cC99b7qGwSNG02LVlgC8/uor3Hr9nygrK6PFqqtyQ997MkgtLeh/kyZw45WXMPWbyQTBbt33Z58DDgPgiUED+fuQh2nQoAFbbLsDR510JpMmjOe0o35NhzXWAmD9jTfl5F4XZfkUpB+1757daNKkKQ2KiiguKuKvAx7lxuv+zPPPjaRhw4asvvoaXHLZH1mlefOso9YLlvt6Yt8e+3PIYUfQ+3fnV+67q38/ttl2O3573Anc1b8fd93ZjzN7nZthSunH7dF9P3oceChXXbpgsZk0cQLjXnqBtqu1r9w3/btp3Hj1FVx1Y1/ardaeb6ZMXt5xpR/VoKiIo08+i3XX34jSGd9z9omHs9mW2zL1m8m8PHYU1/cfSMOVVmLqN1Mq79Ouw+pc339ghqml6unb/15Wbdmycnubbbty6ulnUVxczM3XX8M9d/bjtLPOyTBh/VFry3IiYsOIOD8ibir8nB8RG9XW42nxtthyq4Vm5UeNfIZ99usBwD779WDks09nEU2qUpefbUnz5gu/q3Tb9VdzYs9eRETlvmdGDGeHXbrRrlD4W5a0Wm45pcUpadWGddev+GuwcZOmrL5mJyb/bxJPPvYo+x92DA1XWgmAVVuWZBlTWia27bo9xcUVc8ibdO7CxEkTM05Uf9RKuY+I84GBQAAvF34CGBARF9TGY6rmJk+eTJs2bQFo3boNkyc7w6kVx9jnnqV1m7asu/4GC+z//L+fMX3aNM46+RhO/M1BPDX88YwSSos2acJ4PvnofdbfaBPGf/EZ77z5Gued/BsuOuM4Pnzv7fmO+5Jexx/KRWccxztvvpZhYmnRgqDnScdy5CG/ZtCjDy80/viQQXTd/ucZJKufamtZzrHAT1NKc+bfGRHXAW8DV/3YnSLiBOAEgJtvu51jjzuhluLphyJigdlPqS6bObOUB+7tz9U33b7QWHl5GR+89y7X3HoHs2fNouexR7DRJp1ZY821l39Q6UeUls7gT5ecw29PPZsmTZtRXl7O9O+m8afb7uXD997mmkvPp++DQ2lZ0pp+A4fTvMWqfPz+O1x58dncdPcjNGnaLOunIC3gjnseoG27dkyZPJmeJx3L2p06sfkWWwFw1x19KS4qYs+998k4Zf1RW+V+LtAB+OwH+9sXxn5USqkf0A+gdA6plrKpoFWrVnz99STatGnL119PoqTEt4K1Yhj/xedMGP8lxx9xAABfT5rIib85iNvuHkCbtu1o0WJVGjduQuPGTej8sy34+MP3LfeqE8rK5nD1Jeew4y/2YrsduwHQuk1btv35rkQE62+0CdGgAdO+nUqLVVtWLtVZd4ONWa3D6oz/4r/8ZIONs3wK0kLatmsHQEmrVuy86y94+99vsfkWWzH0scGMGT2K2/rd7QTiclRba+7PBJ6JiL9HRL/Cz5PAM8AZtfSYqqGddt6VoY8NAWDoY0PYeZduGSeSqmedn6zPoCefY8CQEQwYMoI2bdtx+18fpqRVa7bfcVfeeuNflJeVMXNmKe++/RZrrb1O1pElUkrcevVlrL5WJ/Y76IjK/VvvsAtv/WscAF9+/hllc+bQvMWqfDv1G8rLywGYMP4Lvvryv7Rr3zGT7NKilM6Ywffff195+58vjmXdn6zHC2Of57577uTaG2+jUePGGaesXyKl2pkgj4gGwNbAvD+JvgReSSmVV+f+ztwvWxec24txr7zM1KnfUNKqFSefchq7dPsF5519Jl999RUdOnTg6mtvoEWLVbOOmitTvp+ddYRcuLz3ebzx2it8O3UqLUtKOPqEU9lr3/0rxw/tsQd97xlYeSnMgffdzYhhQ4gGDdhr3/054NAjs4qeG9/OmFP1QVqsd976FxedfixrrfMTKv6KhCOO60nnLbbhlqt/zycffUDDhg056qQz6bz51rz43DMMuPsvFBUX06BBAw45+kS26rpTxs9ixbZ6iSVzWfvii88576zTACgrK+OXe3Xnt8efxK+678Hs2bNpsWpFr9h00y5cePHvM0yaL80bNVjkWyG1Vu6XluVeeWC5V15Y7pUHlnvlxeLKvd9QK0mSJOWE5V6SJEnKCcu9JEmSlBOWe0mSJCknLPeSJElSTljuJUmSpJyw3EuSJEk5YbmXJEmScsJyL0mSJOWE5V6SJEnKCcu9JEmSlBOWe0mSJCknLPeSJElSTljuJUmSpJyw3EuSJEk5Uby4wYgoWdx4SmnKso0jSZIkaUktttwDrwIJCGBN4JvC7VWB/wKdajWdJEmSpGpb7LKclFKnlNI6wNPAPiml1imlVkB34KnlEVCSJElS9VR3zf22KaXh8zZSSn8HutZOJEmSJElLoqplOfOMj4jewP2F7cOB8bUTSZIkSdKSqO7M/aFAG2AwMKhw+9DaCiVJkiSp5qo1c1+4Ks4ZEdE0pfR9LWeSJEmStASqNXMfEV0j4h3g3cJ2l4i4rVaTSZIkSaqR6i7LuR7YA5gMkFJ6A9ixtkJJkiRJqrlqf0NtSunzH+wqX8ZZJEmSJC2F6l4t5/OI6AqkiGgInEFhiY4kSZKkuqG6M/cnAacCHYEvgc2AU2orlCRJkqSaq+7M/QYppcPn3xER2wNjl30kSZIkSUuiujP3N1dznyRJkqSMLHbmPiK2A7oCbSKi13xDzYGi2gwmSZIkqWaqWpazEtCscNwq8+2fBhxQW6EkSZIk1VyklKo+KGKtlNJnyyFPpdI5VB1MquOmfD876wjSMvHtjDlZR5CW2uoljbOOIC0TzRs1iEWNVXfNff+IWHXeRkS0jIgRS51MkiRJ0jJT3XLfOqU0dd5GSukboG3tRJIkSZK0JKpb7udGxJrzNiJiLXDZjCRJklSXVPc69xcBYyLiOSCAnwMn1FoqKSeaN6rubzGpbvvJLr2qPkiq4z4ZdX3WEaRlonmjRc/PV6t5pJSejIjNgW0Lu85MKf1vGWSTJEmStIwsdllORGxY+HVzYE1gfOFnzcI+SZIkSXVEVTP3ZwPHA9f+yFgCdl3miSRJkiQtkcWW+5TS8YVfd1k+cSRJkiQtqcWW+4jYf3HjKaVByzaOJEmSpCVV1bKcfQq/tgW6As8WtncBXgAs95IkSVIdUdWynGMAIuIpYOOU0leF7fbAPbWeTpIkSVK1VfdLrNaYV+wLJlJx9RxJkiRJdUR1v2HnmYgYAQwobB8MPF07kSRJkiQtiep+iVXPiPgVsGNhV7+U0uDaiyVJkiSppqo7cw/wGvBdSunpiGgSEauklL6rrWCSJEmSaqZaa+4j4njgUeD2wq6OwJDaCiVJkiSp5qr7gdpTge2BaQAppQ+puDymJEmSpDqiuuV+Vkpp9ryNiCgGUu1EkiRJkrQkqlvun4uI3wGNI2I34BFgaO3FkiRJklRT1S335wNfA28BJwLDgd61FUqSJElSzVV5tZyIKALeTiltCNxR+5EkSZIkLYkqZ+5TSuXA+xHhN9JKkiRJdVh1r3PfEng7Il4Gvp+3M6W0b62kkiRJklRj1S33F9dqCkmSJElLbbHlPiIaAScBP6Hiw7R3ppTKlkcwSZIkSTVT1Zr7e4EtqSj2ewLX1noiSZIkSUukqmU5G6eUNgWIiDuBl2s/kiRJkqQlUdXM/Zx5N1yOI0mSJNVtVc3cd4mIaYXbQcU31E4r3E4ppea1mk6SJElStS223KeUipZXEEmSJElLp8ovsZIkSZK0YrDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlRnHUALR99el/I6NGjKClpxd+GDAPgumv+xOjnRtKwuCGrr7Eml15xJc2bN884qVS18vJyjjz0QNq2bcsNt/TluKOPYMaM7wGYMmUyP92kM9fecEvGKSXo2+dw9txxE76e8h1bHvhHAC46cS9+u39Xvv5mOgB9bnmcEWPeoaRFUx7887Fs8dO1uP/xf3LWnx5Z6HyP3HAinTq2qjyXlIWrLu/Ni2NG07JlCfcMHALAX266hheef47ihsV06LgGF1xyBaus0pyysjlcfUUfPnj/XcrLy9hjr3054ujjM34G+ebMfT2xb4/9ua1v/wX2bbvd9jw6eBiPDB7KWmuvzV39b88onVQzAx64j07rrFO53f+e+3nw4cE8+PBgNu28Gbt0+0WG6aT/d9/Qf7LfqbcutP/m+0ey7SFXse0hVzFizDsAzJw1h8tuG8aF1w/+0XPtt2sXvp8xq1bzStWx5949+PONfRfYt+XW23H3gMHc/eBg1lhzbR64p6JzjHz6KebMmc09AwZzx18fZujgR/hq/JdZxK43aq3cR8SGEdEtIpr9YP8va+sxtWhbbLkVzVu0WGBf1+13oLi44s2bzp03Y+LECVlEk2pk4sQJjH3+OXr86oCFxqZPn864l19i510s96obxr72MVO+nVGtY2fMnM0Lr/+HmbPmLDTWtPFKnH7ErlzV/8llHVGqsS6bb8kqzRfsFFttu31lp9h4k858PWkiABFBaWkpZWVlzJo5i+LihjRt2myhc2rZqZVyHxGnA48BpwH/joj95hv2vcQ6aMjgv7HDDjtmHUOq0rVXX8npZ51DNFj4j69RI59mq222pVkz/+JQ3XbSITvy8kMX0rfP4ay6SuMqj+9zSnduvO8ZZpTOXg7ppKUzfOhgtum6AwA7d9uNxo0bs/9eu3DQvrtx8BFHLzTZqGWrtmbujwe2SCn1AHYGLo6IMwpjsag7RcQJETEuIsbd2b9fLUXTD91x+18oKipir+77Zh1FWqznnxtJSUkJG2380x8df+rvw9ljz72XcyqpZu545Hk23uf3bHPIVUz43zSu6rX/Yo/vvH5HOq3RhsdHvrmcEkpL7r67bqeoqIjdftkdgHfffosGDYoYNPxZBg55kocfuJfxX36eccp8q60P1DZIKU0HSCl9GhE7A49GxFosptynlPoB/QBK55BqKZvm89iQQTw/ehS397+HiEX+r5HqhDde/xejR41k7JjRzJ41m+nfT+fiC8/j8iuvZuo33/D2v9/kz9ffnHVMabEmTfmu8vZdg8Yy6KaTFnv8Nl06scXGa/LeE5dSXNSANiWrMOKOM9jj+BtrO6pUI38fNoQXxozm+tv6V3aKp0cMZ+vttqe4uCEtS1qxSZfNeO+dt+nQcY2M0+ZXbc3cT4yIzeZtFIp+d6A1sGktPaZqaOyY0dx7V39uuPkvNG5c9dvCUtZ6ntGL4f8YxdC/P8Mf/nQtW221DZdfeTUAT/9jBDvsuDMrr7xyximlxVut9f9flWy/XbvwzsdfLfb4Ox4Zwzq7X8SGe/dh12Ou58PPJlnsVee89OIYBtx3F1deezONGv1/p2jXrj2vjXsZgNLSGbzz7zdZa+1OWcWsF2pr5v43QNn8O1JKZcBvIsJLsmTggnN7Me6Vl5k69Rt277YjJ59yGnf178fs2bM56fhjAOjcuQu9+1yWcVJpyTw1YjhH/9bLq6luuffKo/n5FuvRetVmfPTk5Vzedzg7brEenTdYnZQSn301hdOuGFB5/HtPXMoqTRuxUsNi9tmlM91PuZX3/uPFDlS3XNr7XF5/9RW+nTqVA7p345jjT+GBe/sze/Zszu5Z8efwxpt05uwL+9DjwEO56rLeHHXwfiQSe3bvwbrrbZDxM8i3SKlurn5xWY7yoKx8btYRpGWi7XanZx1BWmqfjLo+6wjSMrFai4aLXEvtde4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5USklLLO8KNmllE3g0k1UEd/e7iHUUAAAAt+SURBVEk1NuHbmVlHkJbay59PzjqCtEwc/LOOsagxZ+4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5YTlXpIkScoJy70kSZKUE5Z7SZIkKScs95IkSVJOWO4lSZKknLDcS5IkSTlhuZckSZJywnIvSZIk5URx1gG0/M2aNYtjfnM4c2bPpqy8nN1234NTep6edSypWvr0vpDRo0dRUtKKvw0ZBsCtN9/AqGefIRo0oKSkFZf94Uratm2XcVJpQdf98RJeGjuaVVuWcPv9gwD4z4fvc9Ofr2Bm6Qzate/AeX2upGnTZjw74gkeffDeyvt+8vEH3HLXQNZdf8Os4kuVrut5KCs1bkKDBg1oUFTESX/sy4j7+/L+ay9SVNyQknbt6XHS+TRu2ozysjIe63cN4z/5kLnl5Wy24+7s2OOwrJ9CrkVKKesMP2pmGXUzWA6klCidMYMmTZsyZ84cjj7yMM6/8CI6d9ks62i5U0d/e63QXh33Ck2aNKH3786vLPfTp0+nWbNmADx4/1/5z8cf0bvPZVnGzJ0J387MOsIK763XX6VR4yZcc/lFleX+tGMP4/ievej8sy0ZMWwwE8Z/yVEn9Fzgfp98/CGXXXAmdz/yRBaxc+XlzydnHSEXrut5KCf+sS9Nm7eo3PfRG6/QaZPNKSoq4qkH+gGw++En8OaYZ3jv1Rc46IyLmT1rJrecfQzHXHI9LduullX8XDj4Zx1jUWMuy6mHIoImTZsCUFZWRllZGcQiXyNSnbLFllvRvEWLBfbNK/YApaWlhK9n1UGbbrYFqzRvvsC+Lz//jE032wKAzbfajrHPPbPQ/Ub94+/s9ItfLpeM0pL6SZetKCoqAmD19TZi2pSvKwYCZs8qpby8nLLZsygqbsjKTZpkmDT/lnu5j4hjlvdjamHl5eUctP9+7PLzrmy7XVc6d+6SdSRpqdx84/Xs0W0nhj8xlJN7npF1HKla1uq0Li8+PxKA0SOf4uuJExY6ZvQzI9h5N8u96pAI/vrHc/nLhScy7ulhCw2/NurvrLfZ1gD8dJudWGnlxvz5pAO4tuehbN/9IJo0a77QfbTsZDFzf+miBiLihIgYFxHj7ryj3/LMVO8UFRXx8KDHeOrZ5/j3W2/y4YcfZB1JWiqnnXEWI555jr323oeBD96fdRypWnr97lKGDXqInr89hNIZMyhu2HCB8ffefpOVGzVi7XXWyyihtLDjLr2Rk6/qx5EXXMVLTw3h03ffqBx7bvD9FBUV0XmHXwDwxcfv0aBBA879yyOcddMDjH3iYaZMHJ9V9HqhVj5QGxFvLmoIWOSn3FJK/YB+4Jr75aV58+ZstfU2vDDmedZbb/2s40hLba/u+9Dz5BP8kLhWCGus1Yk/3nA7AF/891NefmH0AuPPPT2CnX+xZxbRpEVqXtIGgGYtWrLRVjvwxUfvsfZGXfjXqCd5/7V/cnTvayqXR7419pmKJTvFxTRr0ZI1N9iE8f/5gJJ2HbJ8CrlWWzP37YDfAPv8yI+fZsnYlClTmDZtGgAzZ87kny++wNqd1sk4lbTkPvvs08rbo559hk6+nrWCmPpNxV+Jc+fOZcC9d7B3jwMrx+bOncvoZ0e43l51yuyZpcwqnVF5++M3x9FujU58+PrLjBn6EIefewUrrdyo8vgWrdryydv/qjz+iw/fpXWHNTLJXl/U1qUwhwHNUkqv/3AgIkbV0mOqmv739SR6/+4C5s4tZ+7cxO57/JKddt4l61hStVxwbi/GvfIyU6d+w+7dduTkU05jzPOj+fTTT2gQQfsOHbnokkWu/pMyc2Wf83nzX+OYNnUqR/TYjSOOPZmZpaUMHTQQgO136sbue/eoPP6t11+lTdvVaN9x9awiSwuZ/u03DLj2EgDmzi2n8/bdWG+zrbnhjCMomzOHe/9wLgCrr7cx+x53Flvv0YMhf/kTN59zDCT42c57sNpa62b5FHLPS2FKtaiO/vaSasxLYSoPvBSm8sJLYUqSJEn1gOVekiRJygnLvSRJkpQTlntJkiQpJyz3kiRJUk5Y7iVJkqScsNxLkiRJOWG5lyRJknLCci9JkiTlhOVekiRJygnLvSRJkpQTlntJkiQpJyz3kiRJUk5Y7iVJkqScsNxLkiRJOWG5lyRJknLCci9JkiTlhOVekiRJygnLvSRJkpQTlntJkiQpJyz3kiRJUk5Y7iVJkqScsNxLkiRJOWG5lyRJknLCci9JkiTlhOVekiRJygnLvSRJkpQTlntJkiQpJyz3kiRJUk5Y7iVJkqScsNxLkiRJOWG5lyRJknLCci9JkiTlhOVekiRJygnLvSRJkpQTlntJkiQpJyKllHUGZSQiTkgp9cs6h7S0fC0rL3wtKw98HWfLmfv67YSsA0jLiK9l5YWvZeWBr+MMWe4lSZKknLDcS5IkSTlhua/fXA+nvPC1rLzwtaw88HWcIT9QK0mSJOWEM/eSJElSTlju66mI2DAiXoyIWRFxTtZ5pCUVEb+MiPcj4qOIuCDrPFJNRcQaETEyIt6JiLcj4oysM0lLIiLuiohJEfHvrLPUZy7Lqacioi2wFtAD+CaldE3GkaQai4gi4ANgN+AL4BXg0JTSO5kGk2ogItoD7VNKr0XEKsCrQA9fx1rRRMSOwHTgrymlTbLOU185c19PpZQmpZReAeZknUVaClsDH6WU/pNSmg0MBPbLOJNUIymlr1JKrxVufwe8C3TMNpVUcyml0cCUrHPUd5Z7SSuyjsDn821/gaVIK7CIWBv4GfBStkkkrags95Ik1QER0Qz4G3BmSmla1nkkrZgs9/VIRJwaEa8XfjpknUdaBr4E1phve/XCPmmFEhENqSj2D6SUBmWdR9KKy3Jfj6SUbk0pbVb4GZ91HmkZeAVYLyI6RcRKwCHA4xlnkmokIgK4E3g3pXRd1nkkrdi8Wk49FRGrAeOA5sBcKj7dvrFvBWtFExF7ATcARcBdKaU/ZBxJqpGI2AF4HniLij+PAX6XUhqeXSqp5iJiALAz0BqYCPRJKd2Zaah6yHIvSZIk5YTLciRJkqScsNxLkiRJOWG5lyRJknLCci9JkiTlhOVekiRJygnLvSTVMxHRIyJSRGxYxXFnRkSTpXicoyPiliW9vySp5iz3klT/HAqMKfy6OGcCS1zuJUnLn+VekuqRiGgG7AAcS8U3+hIRRRFxTUT8OyLejIjTIuJ0oAMwMiJGFo6bPt95DoiIewq394mIlyLiXxHxdES0W97PS5JUoTjrAJKk5Wo/4MmU0gcRMTkitgC2BtYGNksplUVESUppSkT0AnZJKf2vinOOAbZNKaWIOA44Dzi7Np+EJOnHWe4lqX45FLixcHtgYbsT0DelVAaQUppSw3OuDjwUEe2BlYBPllFWSVINWe4lqZ6IiBJgV2DTiEhAEZCAV6p5ijTf7Ubz3b4ZuC6l9HhE7Az8funTSpKWhGvuJan+OAC4L6W0Vkpp7ZTSGlTMsr8BnBgRxVD5jwCA74BV5rv/xIjYKCIaAL+ab38L4MvC7aNq9RlIkhbLci9J9cehwOAf7Psb0B74L/BmRLwBHFYY6wc8Oe8DtcAFwDDgBeCr+c7xe+CRiHgVqGp9viSpFkVKqeqjJEmSJNV5ztxLkiRJOWG5lyRJknLCci9JkiTlhOVekiRJygnLvSRJkpQTlntJkiQpJyz3kiRJUk5Y7iVJkqSc+D9c6/FFVUhPmgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 921.6x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4n1UuTSW6Uqt",
        "colab_type": "text"
      },
      "source": [
        "#Results and ending comet experiment"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pAzaq4Jp8z6l",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "bb56a910-636d-4df4-9976-d17a99559ef5"
      },
      "source": [
        "# import comet_ml in the top of your file\n",
        "from comet_ml import Experiment\n",
        "    \n",
        "# Add the following code anywhere in your machine learning file\n",
        "experiment = Experiment(api_key=\"kyaDe1YHDUV60KbpzF3dVpIuk\",\n",
        "                        project_name=\"general\", workspace=\"rachel-ramonyai\")"
      ],
      "execution_count": 109,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "COMET INFO: Experiment is live on comet.ml https://www.comet.ml/rachel-ramonyai/general/17c090bb37fb4a8a85a802e4518f3ab9\n",
            "\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M467uMBl6Xzt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "f1 = f1_score(y_test, y_predsm,average='macro')\n",
        "precision = precision_score(y_test, y_pred,average='macro')\n",
        "recall = recall_score(y_test, y_pred,average='macro')"
      ],
      "execution_count": 110,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5bxLlvG46axx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "params = {\"kernel\": 'linear',\n",
        "          \"model_type\": \"SVC\",\n",
        "          \"stratify\": True\n",
        "          }"
      ],
      "execution_count": 111,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u68WdakU6fA3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "params = {\n",
        "          \"model_type\": \"Best LinearSVC\",\n",
        "          \"stratify\": True\n",
        "          }\n",
        "\n",
        "metrics = {\"f1\": f1,\n",
        "           \"recall\": recall,\n",
        "           \"precision\": precision\n",
        "           }"
      ],
      "execution_count": 112,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TaeNwjyg6id1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Log our parameters and results\n",
        "experiment.log_parameters(params)\n",
        "experiment.log_metrics(metrics)"
      ],
      "execution_count": 113,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8XG_OiFo6pgh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 433
        },
        "outputId": "2520eb60-7136-43fa-b660-9d8d6fc861f7"
      },
      "source": [
        "experiment.end()"
      ],
      "execution_count": 114,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "COMET INFO: ---------------------------\n",
            "COMET INFO: Comet.ml Experiment Summary\n",
            "COMET INFO: ---------------------------\n",
            "COMET INFO:   Data:\n",
            "COMET INFO:     display_summary_level : 1\n",
            "COMET INFO:     url                   : https://www.comet.ml/rachel-ramonyai/general/17c090bb37fb4a8a85a802e4518f3ab9\n",
            "COMET INFO:   Metrics:\n",
            "COMET INFO:     f1        : 0.597348858332672\n",
            "COMET INFO:     precision : 0.7185310754316903\n",
            "COMET INFO:     recall    : 0.5647107964054541\n",
            "COMET INFO:   Parameters:\n",
            "COMET INFO:     model_type : Best LinearSVC\n",
            "COMET INFO:     stratify   : True\n",
            "COMET INFO:   Uploads:\n",
            "COMET INFO:     code                : 1 (20 KB)\n",
            "COMET INFO:     environment details : 1\n",
            "COMET INFO:     filename            : 1\n",
            "COMET INFO:     installed packages  : 1\n",
            "COMET INFO:     notebook            : 1\n",
            "COMET INFO:     os packages         : 1\n",
            "COMET INFO: ---------------------------\n",
            "COMET INFO: Uploading stats to Comet before program termination (may take several seconds)\n",
            "COMET INFO: Waiting for completion of the file uploads (may take several seconds)\n",
            "COMET INFO: Still uploading\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "DQIx8LdJ5XkN"
      },
      "source": [
        "#Model Analysis and Insights"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "yBIirzqoqHg1"
      },
      "source": [
        "We worked with various models, and at first the linear SV model and the SGD classifier performed the best, with around 70% accuracy. \n",
        "\n",
        "We then decided to fit these two models and increase their performance. Both of the models achieved different precision,recall & F1 scores.\n",
        "\n",
        "With the SV model performing better with precision, but acheiving an accuarcy of 72%. While the SGD model achieved better scores in the other categories, however both models achieved the same accuracy at first.\n",
        "\n",
        "The last model we made after extensive cleaning achieved a lower accuracy by 1%, but the predictions between anti,pro,neutral and news were closer together. Without a clear leader. This better reflects the real world findings.\n",
        "\n",
        "\n",
        "The 'Pro' climate change class accounts for 54% of the data set and the remaining 46% is shared amongst the three other classes. \n",
        "\n",
        "This 54% count for 'Pro' class means more people agree that climate change is a real threat to our eco-system. Companies that offer products and services which are environmentally friendly and sustainable are more likely to receive support from people based on the shared sentiments and beliefs about climate change. \n",
        "\n",
        "The market reasearch done by these companies has positive results which is an indication for prospects of goood profit margins for eco-friendly products and services.\n",
        "\n",
        "Accuracy on holdout data set is always lower and this shows that our model tends to be slightly overfitting the training data...best performing model is linear svc.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "AAuPQ7QL562Y"
      },
      "source": [
        "#Conclusion"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "07ia2EcNx7bE"
      },
      "source": [
        "Our model predicted the news category best, while the other categories did improve, with our model being able to predict anti sentiment the second best. Although only more accurate by a couple of percent. So overall the model performed as anticipated.\n",
        "\n",
        "Through the classification techniques, companies will be able to  access a broad base of consumer sentiments, spanning multiple demographic and geographic categories, this will increase their insights and inform  strategies.\n",
        "\n",
        "This includes adding to their market research efforts in gauging how their product or service may be received.\n",
        "\n",
        "Based on the findings it can be concluded that more customers will be receptive to eco-friendly products and services provided by companies. \n",
        "\n",
        "The results prove that more people believe that climate change is a threat to our lives and that it is up to us to reduce our carbon footprint.\n"
      ]
    }
  ]
}