Findings from the Bank Marketing Dataset: Using Machine Learning to Forecast Term Deposit Subscriptions

Shurui Du

doi:10.54254/2754-1169/2025.BL29270

Advances in Economics, Management and Political SciencesOpen access

Findings from the Bank Marketing Dataset: Using Machine Learning to Forecast Term Deposit Subscriptions

Research Article

Open Access

Findings from the Bank Marketing Dataset: Using Machine Learning to Forecast Term Deposit Subscriptions

Shurui Du ^1*

¹ Boston University

^*Corresponding author: shuruidu@bu.edu

Published on 11 November 2025

AEMPS Vol.240

ISSN (Print): 2754-1177

ISSN (Online): 2754-1169

ISBN (Print): 978-1-80590-527-1

ISBN (Online): 978-1-80590-528-8

Download Cover

Abstract

Banks face the challenge of determining which customers are most likely to react effectively to marketing initiatives in a financial climate that is becoming more and more competitive. Using a large dataset of campaign-related, financial, and demographic data, this study applies machine learning techniques to forecast term deposit subscriptions. To compare many classifiers based on various performance metrics, the study uses a methodical procedure that includes data pretreatment, exploratory data analysis, model training, and evaluation. The findings show that, in comparison to traditional models like logistic regression and decision trees, ensemble approaches—in particular, gradient boosting—achieve improved prediction accuracy and generalization. Gradient Boosting is the most effective classifier for reducing class imbalance in subscription prediction since it attains the biggest area under the Receiver Operating Characteristic (ROC) curve while maintaining a fair balance between precision and recall. These findings demonstrate how targeted marketing campaigns in the banking industry could benefit from sophisticated predictive modeling. Machine learning models maximize marketing return on investment, decrease needless interactions, and improve customer happiness by more accurately identifying high-potential clients. The study emphasizes how crucial it is for financial marketing strategy to include data-driven decision-making as a fundamental element.

Keywords:

Predictive modeling, machine learning, financial marketing strategy, customer segmentation, precision-recall

View PDF

References

[1]. Yu, Q. et al. (2025) Enhancing Bank Term Deposit Predictions: A Machine Learning Approach With CatBoost and SHAP. Applied and Computational Engineering, 120, 171-180.

[2]. Moro, S., Cortez, P. & Rita, P. (2014) A Data-Driven Approach to Predict the Success of Bank Telemarketing. Decision Support Systems, 62, 22-31.

[3]. UCI Machine Learning Repository. (2024) Bank Marketing Data Set. Retrieved from https: //archive.ics.uci.edu/dataset/222/bank+marketing

[4]. Peter, M. et al. (2025) Predicting Customer Subscription in Bank Telemarketing Campaigns Using Hybrid Ensemble Models. Journal of Data Science and Artificial Intelligence, 3(1), 15-28.

[5]. Tanvir, M. F., Hossain, M. & Asifuzzaman, J. (2024) Bayesian Regression for Predicting Subscription to Bank Term Deposits in Direct Marketing Campaigns. ArXiv Preprint, arXiv: 2410.21539.

[6]. Breiman, L. (2001) Random Forests. Machine Learning, 45(1), 5-32.

[7]. Friedman, J. H. (2001) Greedy Function Approximation: A Gradient Boosting Machine. Annals of Statistics, 29(5), 1189-1232.

[8]. Chen, W., Zhan, Z. & Wang, J. (2024) A Survey on Imbalanced Learning: Latest Research and Future Directions. Artificial Intelligence Review.

[9]. López-Pinaya, W. H. et al. (2023) Evaluating Classifier Performance With Highly Imbalanced Big Data. Journal of Big Data, 10, 54.