Volume 20, Issue 4 (12-2025)
J. Mon. Ec. 2025, 20(4): 581-596 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Mahdavi G, Heidarzadeh Azar R, Ofoghi R. Predicting Life Insurance Policyholder Churn in Iran Using Machine Learning: A Transparent and Actionable Framework. J. Mon. Ec. 2025; 20 (4) :581-596
URL: http://jme.mbri.ac.ir/article-1-743-en.html
URL: http://jme.mbri.ac.ir/article-1-743-en.html
1- ECO College of Insurance, Allameh Tabataba’i University, Tehran, Iran
2- Allameh Tabataba’i University
2- Allameh Tabataba’i University
Abstract: (618 Views)
This study tackles the challenge of customer churn in life insurance, which leads to substantial financial losses. It introduces a transparent, reproducible, and leakage-free machine learning framework designed to identify at-risk policyholders accurately and efficiently.
Using 20,000 anonymized Iranian life insurance policies with a churn rate of 26%, the study develops a complete modeling pipeline that includes imputation, standardization, and encoding steps performed strictly within the training process to prevent data leakage. Three model types—logistic regression, random forest, and XGBoost—were trained and evaluated using stratified cross-validation, F1-optimized thresholds, and performance metrics such as ROC-AUC, F1, and precision.
Results show that all models perform similarly (ROC-AUC ≈ 0.70), with logistic regression achieving the highest F1 score (0.66) and XGBoost offering the best precision (0.68). The top-ranked predictions captured about 69% of churners, demonstrating strong operational potential. Policyholders with major payment delays were identified as the most churn-prone and easily detectable group. Overall, the findings confirm that transparent and interpretable machine learning models can effectively balance accuracy, simplicity, and practicality—supporting data-driven, value-focused customer retention strategies in the life insurance industry
Using 20,000 anonymized Iranian life insurance policies with a churn rate of 26%, the study develops a complete modeling pipeline that includes imputation, standardization, and encoding steps performed strictly within the training process to prevent data leakage. Three model types—logistic regression, random forest, and XGBoost—were trained and evaluated using stratified cross-validation, F1-optimized thresholds, and performance metrics such as ROC-AUC, F1, and precision.
Results show that all models perform similarly (ROC-AUC ≈ 0.70), with logistic regression achieving the highest F1 score (0.66) and XGBoost offering the best precision (0.68). The top-ranked predictions captured about 69% of churners, demonstrating strong operational potential. Policyholders with major payment delays were identified as the most churn-prone and easily detectable group. Overall, the findings confirm that transparent and interpretable machine learning models can effectively balance accuracy, simplicity, and practicality—supporting data-driven, value-focused customer retention strategies in the life insurance industry
Keywords: Customer churn prediction, Life insurance, Machine learning, Data leakage prevention, Retention strategy
Type of Study: Original Research - Empirical |
Subject:
Microeconomics
Received: 8 Oct 2025 | Accepted: 18 Nov 2025 | Published: 23 Nov 2025
Received: 8 Oct 2025 | Accepted: 18 Nov 2025 | Published: 23 Nov 2025
References
1. Adadi, A., & Berrada, M. (2018). Peeking inside the black-box: A survey on explainable AI. IEEE access, 6, 52138-52160. [DOI:10.1109/ACCESS.2018.2870052]
2. Alpaydin, E. (2020). Introduction to machine learning (4th ed.). MIT Press.
3. Bansal, H. S., Taylor, S. F., & St. James, Y. (2005). "Migrating" to new service providers: Toward a unifying framework of consumers' switching behaviors. Journal of the academy of marketing science, 33 (1), 96-115. [DOI:10.1177/0092070304267928]
4. Bhattacherjee, A. (2001). Understanding information systems continuance: An expectation-confirmation model 1. MIS Quarterly, 25 (3), 351-370. [DOI:10.2307/3250921]
5. Dietterich, T. G. (2000). Ensemble methods in machine learning. In International workshop on multiple classifier systems (pp. 1-15). Berlin, Heidelberg: Springer Berlin Heidelberg. [DOI:10.1007/3-540-45014-9_1]
6. Geiler, L., Affeldt, S., & Nadif, M. )2022(. Machine learning methods for churn prediction: A survey. International Journal of Data Science and Analytics, 14 (3), 217-242. [DOI:10.1007/s41060-022-00312-5]
7. Groll, A., Wasserfuhr, C., & Zeldin, L. )2022(. Churn modeling of life insurance policies via statistical and machine learning methods: Analysis of important features. arXiv preprint arXiv:2202.09182.
8. Hanafy, M., & Ming, R. )2021). Machine learning approaches for handling auto insurance big data. Risks, 9 (2), 42. [DOI:10.3390/risks9020042]
9. Imani, M., & Arabnia, H. R. (2023). Hyperparameter optimization and sampling techniques for customer churn prediction. Technologies, 11 (6), 167. [DOI:10.3390/technologies11060167]
10. Kotler, P., & Keller, K. L. (2016). Marketing management (15th ed.). Pearson.
11. Lalwani, P., Mishra, M. K., Chadha, J. S., & Sethi, P. (2022). Customer churn prediction system: a machine learning approach. Computing, 104(2), 271-294. [DOI:10.1007/s00607-021-00908-y]
12. Manzoor, A., Qureshi, M. A., Kidney, E., & Longo, L. (2024). Machine learning for customer churn prediction: A practitioner-oriented review. IEEE Access, 12, 70434-70449. [DOI:10.1109/ACCESS.2024.3402092]
13. Ngai, E. W. T., Xiu, L., & Chau, D. C. K. (2009). Application of data mining techniques in customer relationship management: A literature review and classification. Expert Systems with Applications, 36 (2), 2592-2602. [DOI:10.1016/j.eswa.2008.02.021]
14. Shahroodi, K., Avakh Darestani, S., Soltani, S., & Eisazadeh Saravani, A. (2024). Developing strategies to retain organizational insurers using a clustering technique: Evidence from the insurance industry. Technological Forecasting and Social Change, 201, 123217. [DOI:10.1016/j.techfore.2024.123217]
15. Vafeiadis, T., Diamantaras, K. I., Sarigiannidis, G., & Chatzisavvas, K. Ch. (2015). A comparison of machine learning techniques for customer churn prediction. Simulation Modelling Practice and Theory, 55, 1-9. [DOI:10.1016/j.simpat.2015.03.003]
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
