В данном репозитории хранятся выполненные мною проекты, в рамках обучения на курсе Яндекс. Практикума "Специалист по Data Science"

A game theoretic approach to explain the output of any machine learning model.

This projects aims to classify potential churn customers using a Telco Customer Dataset from IBM. The main applications are about the explainability integration with SHAP algorithm and the creation of an interactive dashboard with Exploratory, Classification and Explainability insights.

Enhancing Explainability in Fake News Detection uses SHAP and BiLSTM models to improve the transparency and interpretability of detecting fake news, providing insights into the model's decision-making process.

This repository focuses on Explainable AI techniques for explaining IOT attack detection models and image classification models.

Efficient R implementation of SHAP

🔅 Shapash: User-friendly Explainability and Interpretability to Develop Reliable and Transparent Machine Learning Models

Weighted Shapley Values and Weighted Confidence Intervals for Multiple Machine Learning Models and Stacked Ensembles

The Fraud Detection project aims to improve identification of fraudulent activities in e-commerce and banking by developing advanced machine learning models that analyze transaction data, employ feature engineering, and implement real-time monitoring for high accuracy fraud detection.

Python package for AutoML on Tabular Data with Feature Engineering, Hyper-Parameters Tuning, Explanations and Automatic Documentation

Validation (like Recursive Feature Elimination for SHAP) of (multiclass) classifiers & regressors and data used to develop them.

Automated Tool for Optimized Modelling

The project consists of data obtained from the National Science Foundation’s (NSF) National Ecological Observatory Network (NEON) database. This project obtained data sets of quantified variables that are related to surface water quality, identify suitable predictors and a target response to construct ensemble based models.

Random Forest Algorithms to predict climate impact-drivers (CID), a.k.a., climate extreme indices for impact studies, in crop yields of soybean maize using Random Forest and XGBoost in a SHAP (SHapley Additive exPlanations) framework

scripts used for neural decoding of single and multi unit auditory cortex data

Build a Web App called AI-Powered Heart Disease Risk Assessment App

A methodology designed to measure the contribution of the features to the predictive performance of any econometric or machine learning model.

Bachelor thesis regarding bus passenger forecasting for line 1A-6A in Aarhus, Denmark. We evaluate classical naive and statistical models (Lasso) together with machine learning (Random Forest and XGBoost) and deep learning models (RNN and LSTM). We adopt Performance Based Shapley Values for "black-box" model explaination.

Prediction of NYC taxi trip duration using machine learning