Design and Implementation of a Customer Churn Prediction Model for Zimbabwean Banks Using Machine Learning Techniques

Abstract

The loss of customers in the banking industry due to account closures or the termination of banking services is known as customer churn. This could have serious repercussions for banks, such as decreased revenue, increased expenses to draw in new customers, and harm to their reputation. This study explores the problem of customer churn in Zimbabwean banks and looks into the efficacy of traditional machine learning algorithms in order to forecast and analyze customer churn. The study uses a dataset of customer data that includes demographic and banking-related features to train and assess a number of models, including logistic regression, decision trees, random for ests, and XGBoost. The results show that every model was able to accurately forecast customer attrition. The study illustrates the value of anticipating customer attrition in the banking industry and shows how well machine learning algorithms perform in this regard. The study does point out certain limitations, though, such as the dataset's limited feature set and small sample size, which could limit how broadly the results can be applied. The study's conclusion addresses how the results might be applied to banks in Zimbabwe and offers possible avenues for further investigation.