Conclusion
After training and evaluating all three machine learning models on the banking customer dataset, the Random Forest model emerged as the best performing model with the highest accuracy of approximately 88% and the highest AUC score of approximately 0.87.
Random Forest outperformed the other models because it is an ensemble learning method that combines multiple decision trees, which reduces overfitting and handles class imbalance better. It is also capable of capturing complex non-linear relationships between features which are commonly found in banking customer data.
The churn prediction model can help banks identify customers who are at risk of leaving and take proactive retention measures such as offering special discounts, personalized services, or loyalty programs — ultimately reducing customer attrition and increasing profitability.
