Explainable Customer Churn Prediction Using Machine Learning and Intelligent Retention Recommendations

Author

P. Jasmine Lois Ebenezer, P. Abinisha

Abstract

 Customer churn presents a significant obstacle for the telecommunications sector, as the loss of current customers can impact income and sustainable business development. This initiative suggests an interpretable machine learning framework for forecasting customer churn and offering tailored retention strategies. The IBM Telco Customer Churn dataset is used to analyze customer characteristics such as tenure, contract type, monthly charges, internet service and payment method. The data is preprocessed and encoded before training different machine learning models. Support Vector Machine (SVM) is selected as the final prediction model based on its overall performance, achieving an accuracy of 81.41%. To improve the interpretability of the predictions, SHAP (SHapley Additive exPlanations) is used to identify the customer features that contribute to churn predictions. Based on the predicted risk and influential customer characteristics, the system generates personalized retention recommendations. A Streamlit-based application provides customer prediction, churn probability, risk assessment, explanations, recommendations, prediction history and an administrative dashboard. The proposed system aims to support telecommunications businesses in identifying customers at risk of churn and taking suitable preventive actions.

Keywords

Customer Churn, Machine Learning, SVM, Explainable AI, SHAP, Telecommunications, Retention Recommendation.

Full Text:

Download Paper PDF

References

[1] Algorithms paper, 2024, “Customer churn prediction” using ensemble learning and XAI techniques, DOI: 10.3390/a17060231. The paper reports Random Forest as the strongest model among its evaluated approaches and uses LIME and SHAP for explainability.

[2] D. Asif, M. S. Arif, and A. Mukheimer, “A data-driven approach with explainable artificial intelligence for customer churn prediction in the telecommunications industry,” Results in Engineering, vol. 26, 104629, 2025.

[3] M. N. H. Mamun, “Advancements in Machine Learning for Customer Retention: A Systematic Literature Review of Predictive Models and Churn Analysis,” Journal of Sustainable Development and Policy, vol. 1, no. 1, pp. 250–284, 2025, DOI: 10.63125/9b316w70.

Share your valuable work from Social Media Buttons