Development of a Hybrid Fuzzy Logic and SHAP-Based Evaluation Framework for the Prediction of Bank Deposits in the Banking Sector

M. B. Namazov; M. E. Rahimov

doi:10.15802/stp2026/354457

Authors

M. B. Namazov Inst. «Cyber-Physical Systems and Cybersecurity», Azerbaijan https://orcid.org/0000-0003-1119-2442
M. E. Rahimov Dep. «Engineering Mathematics and Artificial Intelligence», Azerbaijan https://orcid.org/0009-0009-0961-5064

DOI:

https://doi.org/10.15802/stp2026/354457

Keywords:

deposit prediction, banking customer behavior, machine learning, fuzzy multi-criteria decision analysis, entropy-based weighting, explainable artificial intelligence, temporal behavior analysis

Abstract

Purpose. This study examines the decision-making behavior of bank customers regarding deposit participation and addresses the challenge of accurately predicting deposit decisions in the presence of class imbalance, limited interpretability, and variability in customer engagement patterns. Methodology. A comparative experimental framework was developed to evaluate several classical and ensemble machine learning approaches using a banking customer dataset. To ensure a balanced and objective comparison of competing models, a hybrid evaluation framework based on fuzzy multi-criteria decision analysis and entropy-based weighting was employed. In addition, explainable artificial intelligence techniques were applied to interpret the contribution of individual variables, while temporal analysis was conducted to investigate behavioral variations in customer response over time. Findings. The results indicate that ensemble learning approaches demonstrate more stable predictive capability compared to conventional models. The hybrid evaluation framework enables consistent ranking of predictive methods and improves the reliability of model selection. Feature interpretation and temporal behavioral analysis reveal that customer interaction characteristics and communication-related variables play a decisive role in deposit participation decisions. Originality. The study proposes an integrated predictive modeling framework combining machine learning forecasting, entropy-weighted fuzzy decision evaluation, explainable artificial intelligence, and time-based behavioral analysis, providing a transparent and robust mechanism for model comparison. Practical value. The proposed approach can assist financial institutions in improving marketing campaign planning, optimizing customer targeting strategies, and supporting data-driven decision-making in deposit acquisition processes

References

Abedin, M. Z., Hajek, P., Sharif, T., Satu, Md. S., & Khan, Md. I. (2023). Modelling bank customer behaviour using feature engineering and classification techniques. Research in International Business and Finance, 65, 101913. DOI: https://doi.org/10.1016/j.ribaf.2023.101913 (in English)

Basten, C., & Mariathasan, M. (2023). Interest rate pass-through and bank risk-taking under negative-rate poli-cies with tiered remuneration of central bank reserves. Journal of Financial Stability, 68, 101160. DOI: https://doi.org/10.1016/j.jfs.2023.101160 (in English)

Bogaard, H., Doerr, S., Jonker, N., Kiefer, H., Koltukcu, O., Lopez, C., Ornelas, J. R. H., Rambharat, R, Röhrs, S., Teppa, F., Bruggen, F. van, & Vansteenberghe, E. (2024). Literature review on financial technology and competition for banking services. Bank for International Settlements, 43, 41. Retrieved from https://www.bis.org/bcbs/publ/wp43.htm (in English)

Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5-32. DOI: https://doi.org/10.1023/A:1010933404324 (in English)

Chen, T., & Guestrin, C. (2016, n.d.). XGBoost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 785-794). New York, NY, USA. DOI: https://doi.org/10.1145/2939672.2939785 (in English)

Davis, J., & Goadrich, M. (2006, 25-29 June). The relationship between precision-recall and ROC curves. In Pro-ceedings of the 23rd International Conference on Machine Learning (pp. 233-240). Pittsburgh, Pennsylvania. DOI: https://doi.org/10.1145/1143844.1143874 (in English)

Fawcett, T. (2006). An introduction to ROC analysis. Pattern Recognition Letters, 27(8), 861-874. DOI: https://doi.org/10.1016/j.patrec.2005.10.010 (in English)

Grandi, P., & Guille, M. (2023). Banks, deposit rigidity and negative rates. Journal of International Money and Finance, 133, 102810. DOI: https://doi.org/10.1016/j.jimonfin.2023.102810 (in English)

Haddadi, S. J., Farshidvard, A., Silva, F. dos S., dos Reis, J. C., & da Silva Reis, M. (2024). Customer churn prediction in imbalanced datasets with resampling methods: A comparative study. Expert Systems with Applications, 246, 123086. DOI: https://doi.org/10.1016/j.eswa.2023.123086 (in English)

Hastie, T., Tibshirani, R., & Friedman, J. (2009). The elements of statistical learning: Data mining, inference, and prediction. Springer New York. DOI: https://doi.org/10.1007/978-0-387-84858-7 (in English)

Kilimci, Z. H. (2022). The effectiveness of homogeneous classifier ensembles on customer churn prediction in banking, insurance, and telecommunication sectors. International Journal of Computational and Experimental Science and Engineering, 8(3), 77-84. DOI: https://doi.org/10.22399/ijcesen.1163929 (in English)

Lundbegr, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. In Advances in Neural Information Processing Systems, 30, 4768-4777. Retrieved from https://papers.nips.cc/paper_files/paper/2017/hash/8a20a8621978632d76c43dfd28b67767-Abstract.html (in English)

Marioriyad, A., & Ramazi, P. (2025). Optimizing accuracy, recall, specificity, and precision using ILP. Mathematics, 13(7), 1059. DOI: https://doi.org/10.3390/math13071059 (in English)

Moro, S., Cortez, P., & Rita, P. (2014). A data-driven approach to predict the success of bank telemarketing. Decision Support Systems, 62, 22-31. DOI: https://doi.org/10.1016/j.dss.2014.03.001 (in English)

Schölkopf, B., Smola, A. J., & Müller, K.-R. (2008). Kernel methods in machine learning. Annals of Statistics, 36(3), 1171-1220. DOI: https://doi.org/10.1214/009053607000000677 (in English)

Shannon, C. E. (1948). A mathematical theory of communication. Bell System Technical Journal, 27(3), 379-423. DOI: https://doi.org/10.1002/j.1538-7305.1948.tb01338.x (in English)

Văduva, A.-G., Oprea, S.-V., Niculae, A.-M., Bâra, A., & Andreescu, A.-I. (2024). Improving churn detection in the banking sector: A machine learning approach with probability calibration techniques. Electronics, 13(22), 4527. DOI: https://doi.org/10.3390/electronics13224527 (in English)

Wieland, F. G., Teuho, J., & Klén, R. (2024). Evaluation metrics and statistical tests for machine learning. Scientific Reports, 14, 56706. DOI: https://doi.org/10.1038/s41598-024-56706-x (in English)

Xie, C., Zhang, J.-L., Zhu, Y., Xiong, B., & Wang, G.-J. (2023). How to improve the success of bank telemarketing? Prediction and interpretability analysis based on machine learning. Computers & Industrial Engineer-ing, 175, 108874. DOI: https://doi.org/10.1016/j.cie.2022.108874 (in English)

Zadeh, L. A. (1965). Fuzzy sets. Information and Control, 8(3), 338-353. DOI: https://doi.org/10.1016/s0019-9958(65)90241-x (in English)