Bank Customer Churn Prediction

Abstract

In the current challenging era, there is a stiff competition happening between the banking industries. To strengthen the grade and level of services they provide, banks focus on customer retention as well as the customer churning. Customer churning becomes one of the duties of corporate intelligences to speculate the number of customers leaving from the bank or presumed to be churned. It also helps in predicting the number of customers retained. The primary objective of this paper is "Bank customer churn prediction" is to build a model that can distinguish and visualize which factors or attributes contribute to customer churn. In addition to that, this paper also discusses a comparison between various classification algorithms. Machine learning is a modern technology that has the potential to solve classification problems. Using supervised machine learning techniques, a best model is chosen that will assign a probability to the churn to simplify customer service to prevent customer churn. Few methodologies are compared in order to accomplish different accuracy levels. XGBoost is considered in order to check if a better model can be obtained that provides best result in terms of accuracy. The other three machine learning algorithms compared are Logistic regression, Support vector machine [SVM], and Random Forest.