Implementation of Machine Learning Techniques for predicting Credit Card Customer action

• Published in: 2022 International Conference on Innovative Computing, Intelligent Communication and Smart Electrical Systems (ICSES) pp. 1 - 5
• Main Authors: Panduro-Ramirez, Jeidy, Akram, Shaik Vaseem, Reddy, Ch.Srinivasa, Ruiz-Salazar, Jenny Maria, Kanwer, Budesh, Singh, Ram
• Format: Conference Proceeding
• Language: English
• Published: IEEE 15-07-2022
• Subjects: Banking Industry; Boosting; Catboost; Churn Prediction; Credit cards; Forestry; Insurance; Machine Learning; Machine learning algorithms; Neural networks; Prediction algorithms; Random forest Classifier; SMOTE
• DOI: 10.1109/ICSES55317.2022.9914238

When a consumer switches from one service provider to another, they are considered a churner. With an expanding number of fierce competitors inside the industry, important banks place a premium on client relationship management. A detailed and real-time credit card holder churn review is critical and helpful for bankers looking to retain credit cards. According to extensive research, maintaining an existing client is more than five times simpler than acquiring a new one. As a result, this research provides a strategy for predicting churns using a bank dataset. The "Synthetic Minority Oversampling Technique" (SMOTE) was employed in this study to handle the unbalanced dataset. Randome forest, K closest neighbour, and two boosting algorithms, XgBoost and CatBoost, are used to forecast credit card user turnover. To improve accuracy, hyperparameter tweaking using grid search was performed. The testing results demonstrate that Catboost has an accuracy of 97.85 percent and outperforms the other models.