Predicting Churning Bank Customers

Customers' churn, which is the common measure of lost customers, is one of the major problems in industries such as banks where there is a fierce competition. By minimizing the number of churning customers companies can maximize their profit and sustainability. This book presents the prediction of bank customers who are prone to move to a competitor. The data of 13172 customers and their corresponding 628,634 transactions is collected from the bank. The CRISP-DM methodology is followed. After the business goals are clearly identified, data preparation processes are undertaken. A data-set of 6045 instances and 18 attributes is prepared. SMOTE (Synthetic Minority Oversampling Technique) has been applied to minimize the class imbalance problem. The modeling techniques applied J48, Logistic Regression, and Bagging. The models are compared by their F-Measure values on separate test sets. The J48 modeling technique gives the best model with a predicting performance of 94.8%. This work is helpful for students, academic or non-academic researchers who are interested in the area of Data Mining specifically in predicting the behavior of churning customers in banks and related industries.