Bank Customer Churn Prediction Using Machine Learning | ||||
| Journal of Engineering Advances and Technologies for Sustainable Applications | ||||
| Volume 1, Issue 3, July 2025, Page 1-8 PDF (862.97 K) | ||||
| Document Type: Original research paper | ||||
| DOI: 10.21608/jeatsa.2025.440922 | ||||
 View on SCiNiTO | ||||
| Authors | ||||
Mohammed Abd Al-Mohsen Ragab1; Eman El Behiry 2
| ||||
| 1Department of Communications and Electronics Engineering, Giza Engineering Institute, Giza, Egypt | ||||
| 2Giza Engineering Institute | ||||
| Abstract | ||||
| This study investigates customer attrition prediction in the banking industry using a comprehensive customer-level dataset from ABC Multinational Bank. By leveraging historical client behavior, we identify critical factors influencing future attrition. To ensure robust and unbiased comparisons, we evaluate the performance of several supervised machine learning algorithms, including random forests, logistic regression, decision trees, and elastic nets, using a standardized cross-validation framework. The results demonstrate that random forests achieve superior predictive accuracy compared to other methods. Our analysis reveals that customers with stronger relationships with the bank, greater utilization of its products and services, and higher loan uptake are significantly less likely to terminate their accounts. These findings underscore the economic relevance of the predictive model and emphasize the importance of targeted upselling and cross-selling strategies to enhance customer retention. This research offers valuable insights for financial institutions aiming to mitigate attrition and optimize long-term client engagement strategies. | ||||
| Keywords | ||||
| Churn; Churn prediction; Financial services Machine learning; Random forests | ||||
|
Statistics Article View: 72 PDF Download: 59 |
||||
