Repurchase Prediction Notebook Summary

The notebook conducts data cleaning and feature engineering on four CSV files: audiograms.csv, calls.csv, contacts.csv, and sale.csv. Following this, it applies two machine learning models, XGBoost and Logistic Regression, for predicting customer repurchases within a specific timeframe. The process involves cross-validation, hyperparameter optimization using Hyperopt, and evaluation metrics like ROC-AUC. Additionally, it visualizes feature importance using SHAP values for XGBoost and coefficient magnitudes for Logistic Regression, providing insights into predictive factors.