Customer Churn Risk Analysis Using a TabNet and SHAP Model in Banking Sector
Abstrak
Customer churn presents a significant challenge for the banking industry, impacting revenue and business sustainability. However, identifying at risk customers often relies on traditional methods that may lack accuracy or interpretability. This study develops a robust and interpretable churn prediction model using TabNet (Tabular Network) and SHAP (SHapley Additive exPlanations). The model was trained on a large-scale dataset of 115,640 banking customers with a significant class imbalance (12.2% churn) without employing synthetic sampling. The experimental results demonstrate that the proposed TabNet model achieved superior performance, yielding an Accuracy of 99,87% and an F1-score of 99,48% for the churn class, significantly outperforming baseline models, including MLP, DNN, and Simple RNN. Furthermore, SHAP analysis successfully identified ‘Balance’, ‘NumComplaints’, ‘NumOfProducts’, and ‘Credit Score’ as the most critical features influencing churn risk. This combined approach provides financial institutions with a highly accurate predictive tool and actionable insight to formulate targeted retention strategies.
Keywords: Customer Churn; TabNet; SHAP; Banking; Model Interpretability
Referensi
A. Li, T. Yang, X. Zhan, Y. Shi, and H. Li, "Utilizing Data Science and AI for Customer Churn Prediction in Marketing," Journal of Theory and Practice of Engineering Science, vol. 4, no. 5,
pp. 72-79, 2024, doi: 10.53469/jtpes.2024.04(05).10.
X. Xiahou and Y. Harada, "B2C E-Commerce Customer Churn Prediction Based on K-Means and SVM," Journal of Theoretical and Applied Electronic Commerce Research, vol. 17, no. 2, pp. 458-475, 2022, doi: 10.3390/jtaer17020024.
P. Kotler and K. L. Keller, Marketing Management, 15th global ed. London: Pearson Education Limited, 2016.
C. Wang, C. Rao, F. Hu, X. Xiao, and M. Goh, "Risk assessment of customer churn in telco using FCLCNN-LSTM model," Expert Systems with Applications, vol. 248, p. 123352, 2024, doi: 10.1016/j.eswa.2024.123352.
L. Zhang and Q. Wei, "Personalized and contextualized data analysis for e-commerce customer retention improvement with Bi-LSTM churn prediction," IEEE Transactions on Consumer Electronics, vol. 70, no. 1, pp. 1-12,
, doi: 10.1109/TCE.2024.3376672.
K. Peng and Y. Peng, "Research on Telecom Customer Churn Prediction Based on GA- XGBoost and SHAP," Journal of Computer and Communications, vol. 10, no. 11, pp. 107-120,
, doi: 10.4236/jcc.2022.1011008.
X. Liu, G. Xia, X. Zhang, W. Ma, and C. Yu, "Customer churn prediction model based on hybrid neural networks," Scientific Reports, vol. 14, p. 30707, 2024, doi: 10.1038/s41598-024-79603-9.
S. Moro, P. Cortez, and P. Rita, "A data-driven approach to predict the success of bank telemarketing," Decision Support Systems, vol. 62,
pp. 22-31, 2014, doi: 10.1016/j.dss.2014.03.001.
S. Moro, P. Cortez, and P. Rita, "Extreme Gradient Boosting Trees with Efficient Bayesian Optimization for Bank Telemarketing Campaigns," Journal of Open Innovation: Technology, Market, and Complexity, 2025 (In Press).
P. P. Singh, F. I. Anik, R. Senapati, A. Sinha,
N. Sakib, and E. Hossain, "Investigating customer churn in banking: a machine learning approach and visualization app for data science and management," Data Science and Management, vol. 7, no. 1, pp. 7-16, 2024, doi:
1016/j.dsm.2023.09.002.
S. Ö. Arik and T. Pfister, "TabNet: Attentive interpretable tabular learning," in Proceedings of
the AAAI Conference on Artificial Intelligence, vol. 35, no. 8, pp. 6679-6687, 2021, doi: 10.1609/aaai.v35i8.16826.
S. M. Lundberg and S.-I. Lee, "A unified approach to interpretable machine learning," in Advances in Neural Information Processing Systems 30 (NIPS 2017), 2017, pp. 4765–4774.
[Online]. Available:
https://arxiv.org/abs/1705.07874.
Q. Shi and Y. Shen, "Explainable customer churn prediction model based on deep learning," Expert Systems with Applications, vol. 248, p. 123416, 2024, doi: 10.1145/3654823.3654875.
J. Kali, "Self-supervised learning for customer churn prediction," M.S. thesis, Tilburg University, Netherlands, 2023. [Online]. Available: https://www.kaggle.com/datasets/blastchar/telco- customer-churn.
S. Pyo and S. O. Choi, "Regional innovation and economic growth: Empirical insights from FGLS, FE-DKSE, and XGBoost-SHAP approach," Journal of Open Innovation: Technology, Market, and Complexity, vol. 11, p. 100524, 2025, doi: 10.1016/j.joitmc.2025.100524.
Y. Sun and H. Ma, "Interpretable analysis of transformer winding vibration characteristics: SHAP and multi-classification feature optimization," International Journal of Electrical Power and Energy Systems, vol. 166, p. 110585, 2025, doi: 10.1016/j.ijepes.2025.110585.
H. Zhou et al., "Application of conversational intelligent reporting system based on artificial intelligence and large language models," Journal of Theory and Practice of Engineering Science, vol. 4, no. 3, pp. 176-182, 2024, doi: 10.53469/jtpes.2024.04(03).16.
S. D. Mfazi, "Bank Customer Churn Dataset," Kaggle, 2023. [Online]. Available: https://www.kaggle.com/datasets/sandiledesmond mfazi/bank-customer-churn.
