Bank Customer Churn Prediction Using a Hybrid Ensemble Soft Voting Approach Based on Tabnet and XGBOOST
Downloads
In an increasingly competitive banking industry, the ability to predict potential customer churn is a strategic factor in maintaining business profitability and sustainability. Churn has a direct impact on a bank’s revenue and operational efficiency; therefore, a prediction model is needed that is not only accurate but also stable and adaptive to variations in customer data. This study proposes a hybrid ensemble soft voting approach based on TabNet and XGBoost to improve the performance and robustness of churn prediction. TabNet, with its sequential attention mechanism, can selectively identify important features, while XGBoost excels at handling nonlinear relationships and controlling overfitting through gradient boosting regularization. The two models are combined using a probability-based soft voting mechanism to produce more balanced and consistent predictions. To address class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) was applied so that the data distribution is more proportional and churn patterns can be better represented. The experimental results show that the proposed approach achieves optimal performance, with an accuracy of 96.74%, precision of 90.09%, recall of 89.53%, and an F1-score of 89.81%. These values indicate that the model is able to maintain a balance between accurate churn detection and the minimization of misclassification. This hybrid ensemble soft voting approach has proven to be superior to single models in terms of predictive stability and generalization capability, making it an effective framework to support data-driven customer retention strategies in the banking sector.
Alizadeh, M., Zadeh, D. S., Moshiri, B., & Montazeri, A. (2023). Development of a customer churn model for banking industry based on hard and soft data fusion. IEEE Access, 11, 29759–29768. https://doi.org/10.1109/ACCESS.2023.3257352
Azim Mim, M., Majadi, N., & Mazumder, P. (2024). A soft voting ensemble learning approach for credit card fraud detection. Heliyon, 10(3). https://doi.org/10.1016/j.heliyon.2024.e25466
Chowdhury, M. N. H., et al. (2025). Deep learning for early detection of chronic kidney disease stages in diabetes patients: A TabNet approach. Artificial Intelligence in Medicine, 166. https://doi.org/10.1016/j.artmed.2025.103153
Das, D. K. (2024). Exploring the symbiotic relationship between digital transformation, infrastructure, service delivery, and governance for smart sustainable cities. Smart Cities, 7(2), 806–835.
Edwine, N., Wang, W., Song, W., & Ssebuggwawo, D. (2022). Detecting the risk of customer churn in telecom sector: A comparative study. Mathematical Problems in Engineering, 2022. https://doi.org/10.1155/2022/8534739
Fares, I. A., & Abd Elaziz, M. (2025). Explainable TabNet transformer-based on Google Vizier optimizer for anomaly intrusion detection system. Knowledge-Based Systems, 316. https://doi.org/10.1016/j.knosys.2025.113351
Galal, M., Rady, S., & Aref, M. (2022). Enhancing customer churn prediction in digital banking using ensemble modeling. In Proceedings of the 4th Novel Intelligent and Leading Emerging Sciences Conference (NILES 2022) (pp. 21–25). IEEE. https://doi.org/10.1109/NILES56402.2022.9942408
Gala, D. M., Pawar, B., Band, G., Dua, P., Mohanty, B. R., & Patil, B. V. (2024). Enhancing predictive accuracy for customer churn in digital banking: A multi-model analysis. In Proceedings of the IEEE Delhi Section Flagship Conference (DELCON 2024). IEEE. https://doi.org/10.1109/DELCON64804.2024.10866346
Imani, M., Joudaki, M., Beikmohammadi, A., & Arabnia, H. R. (2025). Customer churn prediction: A systematic review of recent advances, trends, and challenges in machine learning and deep learning. Machine Learning and Knowledge Extraction, 7(3), 105.
Kanász, R., Drotár, P., Gnip, P., & Zoričák, M. (2024). Clash of titans on imbalanced data: TabNet vs XGBoost. In Proceedings of the IEEE Conference on Artificial Intelligence (CAI 2024) (pp. 320–325). IEEE. https://doi.org/10.1109/CAI59869.2024.00068
Khan, M. I., et al. (2025). XGBoost-TabNet ensemble model for prediction of methylene blue adsorption on activated carbon geopolymer composite. In Proceedings of the International Conference on Innovation in Artificial Intelligence and Internet of Things (AIIT 2025). IEEE. https://doi.org/10.1109/AIIT63112.2025.11082811
de Lima Lemos, R. A., Silva, T. C., & Tabak, B. M. (2022). Propensity to customer churn in a financial institution: A machine learning approach. Neural Computing and Applications, 34(14), 11751–11768. https://doi.org/10.1007/s00521-022-07067-x
Liu, Z. (2023). A new porosity prediction method based on deep learning of TabNet algorithm. In Proceedings of the IEEE International Conference on Electrical Engineering, Big Data and Algorithms (EEBDA 2023) (pp. 1681–1685). IEEE. https://doi.org/10.1109/EEBDA56825.2023.10090680
Liu, Z., Fan, Q., Wang, Z., & Cai, Y. (2023). A novel algorithm for credit default prediction using TabNet. In Proceedings of the International Conference on Electronic Information Engineering and Computer Science (EIECS 2023) (pp. 24–27). IEEE. https://doi.org/10.1109/EIECS59936.2023.10435411
Ljubičić, K., Merćep, A., & Kostanjčar, Z. (2023). Churn prediction methods based on mutual customer interdependence. Journal of Computational Science, 67. https://doi.org/10.1016/j.jocs.2022.101940
Manzoor, A., Qureshi, M. A., Kidney, E., & Longo, L. (2024). A review on machine learning methods for customer churn prediction and recommendations for business practitioners. IEEE Access, 12, 70434–70463. https://doi.org/10.1109/ACCESS.2024.3402092
McDonnell, K., Murphy, F., Sheehan, B., Masello, L., & Castignani, G. (2023). Deep learning in insurance: Accuracy and model interpretability using TabNet. Expert Systems with Applications, 217. https://doi.org/10.1016/j.eswa.2023.119543
Muneer, A., Ali, R. F., Alghamdi, A., Taib, S. M., Almaghthawi, A., & Ghaleb, E. A. A. (2022). Predicting customers churning in banking industry: A machine learning approach. Indonesian Journal of Electrical Engineering and Computer Science, 26(1), 539–549. https://doi.org/10.11591/ijeecs.v26.i1.pp539-549
Nair, A. S., Krishna, A., Gupta, S. T., & Susan, S. (2025). Credit card fraud detection using soft voting ensemble with imbalance treatment. In Proceedings of the International Conference on Intelligent Technologies (pp. 1–5). IEEE. https://doi.org/10.1109/CONIT65521.2025.11167470
Pulkundwar, P., Rudani, K., Rane, O., Shah, C., & Virnodkar, S. (2023). A comparison of machine learning algorithms for customer churn prediction. In Proceedings of the IEEE International Conference on Advances in Science and Technology (ICAST 2023) (pp. 437–442). IEEE. https://doi.org/10.1109/ICAST59062.2023.10455051
Rana, M. S., et al. (2025). AI-driven predictive modeling for banking customer churn: Insights for the US financial sector. Journal of Ecohumanism, 4(1), 3478–3497. https://doi.org/10.62754/joe.v4i1.6188
Salur, M. U., & Aydın, İ. (2022). A soft voting ensemble learning-based approach for multimodal sentiment analysis. Neural Computing and Applications, 34(21), 18391–18406. https://doi.org/10.1007/s00521-022-07451-7
Singh, P. P., Anik, F. I., Senapati, R., Sinha, A., Sakib, N., & Hossain, E. (2024). Investigating customer churn in banking: A machine learning approach and visualization app for data science and management. Data Science and Management, 7(1), 7–16. https://doi.org/10.1016/j.dsm.2023.09.002
Zheng, C., Chen, Y., & Du, X. (2025). A robust soft voting ensemble of the isolation forest model, extended isolation forest model and generalized isolation forest model for multivariate geochemical anomaly recognition. Ore Geology Reviews, 185. https://doi.org/10.1016/j.oregeorev.2025.106787
Zheng, Y., & Zhang, N. (2024). Research on prediction bank credit card user churn based on XGBoost. In Proceedings of the International Conference on Electronic Technology and Information Science (ICETIS 2024) (pp. 690–693). IEEE. https://doi.org/10.1109/ICETIS61828.2024.10593724
Copyright (c) 2026 Mohamad Syazimmi Hersyaputra, Shintami Chusnul Hidayati
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
