Customer Churn Prediction

Source Code and Dataset for Customer Churn Prediction using Explainable AI (XAI) This repository contains the dataset and Python source code used for a comprehensive comparative analysis of Machine Learning algorithms in predicting customer churn within the telecommunication industry. Key Contents: Dataset (telco.csv): A widely used customer churn dataset from the telecommunications sector. Source Code: A Python script implementing five supervised Machine Learning models (CatBoost, XGBoost, Random Forest, Logistic Regression, and KNN) for comparative performance evaluation. Methodology: The code includes steps for data preprocessing (KNN Imputation, Scaling, One-Hot Encoding), addressing class imbalance (SMOTE), and a detailed implementation of Explainable AI (XAI) techniques, namely SHAP (Global Interpretation) and LIME (Local Interpretation), based on the best-performing model (CatBoost). The materials provided enable the full reproducibility of the results, performance comparison, efficiency assessment, and the managerial insights derived from the XAI framework discussed in the corresponding publication.