Enhancement of XG-Boost Using Custom Hyper Parameter Tuning for Bank Churning

Abstract

Bank is an important component of our society which deals with money transaction i.e., lending and deposit of money. Customer churn is termination of business of a customer with the company. Bank customer churn creates an impact on revenue and operational efficiencies of banks, where a customer switches or leaves availing the services of bank. Bank is an important part of our society since it makes money by lending money to others. To understand customer churning behavior it is necessary to retain customers and increase the number of customers. In order to predict the bank customer churning behavior a few algorithms such as XGBoost, CatBoost, AdaBoost, Random Forest, K Near Neighbor, Decision Tree, and Logistic Regression are analyzed. Finally, the best model has been recommended by analyzing the above-mentioned algorithms