A Comparative Study of Machine Learning Techniques for Credit Card Customer Churn Prediction
- Title
- A Comparative Study of Machine Learning Techniques for Credit Card Customer Churn Prediction
- Creator
- Bose A.; Thomas K.T.
- Description
- A customer is a churner when a customer moves from one service provider to another. Nowadays, with an increasing number of severe competition with inside the market, essential banks pay extra interest on customer courting management. A robust and real-time credit card holders churn evaluation is vital and valuable for bankers to preserve credit cardholders. Much research has been observed that retaining an old customer is more than five times easier compared to gaining a new customer. Hence, this paper proposes a method to predict churns based on a bank dataset. In this work, Synthetic Minority Oversampling Technique (SMOTE) has been used for handling the imbalanced dataset. Credit card customer churn is predicted using random forest, k-nearest neighbor, and two boosting algorithms, XGBoost and CatBoost. Hyperparameter tuning using grid search has been used to increase the accuracy. The experimental result shows Catboost has achieved an accuracy of 97.85% and tends to do better than the other models. 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
- Source
- Lecture Notes on Data Engineering and Communications Technologies, Vol-141, pp. 295-307.
- Date
- 2023-01-01
- Publisher
- Springer Science and Business Media Deutschland GmbH
- Subject
- Banking industry; CatBoost; Churn prediction; Machine learning; Random forest classifier; SMOTE
- Coverage
- Bose A., Department of Data Science, Christ University, Bangalore, India; Thomas K.T., Department of Data Science, Christ University, Bangalore, India
- Rights
- Restricted Access
- Relation
- ISSN: 23674512
- Format
- Online
- Language
- English
- Type
- Book chapter
Collection
Citation
Bose A.; Thomas K.T., “A Comparative Study of Machine Learning Techniques for Credit Card Customer Churn Prediction,” CHRIST (Deemed To Be University) Institutional Repository, accessed February 23, 2025, https://archives.christuniversity.in/items/show/18517.