Explainable AI for Customer Churn Prediction in the Energy Sector Using Ensemble Machine Learning Models
DOI:
https://doi.org/10.4108/airo.9813Keywords:
Customer Churn Prediction, Explainable AI (XAI), Machine Learning, Random Forest, XGBoost, LightGBM, SMOTE, SHAP, LIME, Performance Metrics, ROC-AUC, PR-AUCAbstract
Customer churn prediction is crucial for energy providers to preserve revenue and market share in a competitive setting. This research explores implementing explainable AI (XAI) in customer churn prediction with machine learning algorithms as well as interpretability methods. This research utilizes a dataset formed by combining client and price information from Kaggle’s PowerCo dataset, making use of the ‘id’ column as a key. During the data preprocessing section, the data involves extensive preparation, such as dropping unnecessary columns, deleting duplicates, encoding category features, capping outlier values, and imputing missing values. The class imbalance problem was alleviated through making use of the Synthetic Minority Oversampling Technique (SMOTE) to enable robust training of models. The machine learning algorithms Random Forest, XGBoost, and LightGBM were built, implemented, and benchmarked with principal performance metrics like accuracy, precision, recall, F1-score, as well as ROC-AUC. Among the evaluated ensemble models, XGBoost demonstrated the most balanced trade-off between predictive accuracy and interpretability when combined with SHAP and LIME. SHAP (SHapley Additive exPlanations) as well as LIME (Local Interpretable Model-agnostic Explanations) are used for providing global as well as localinterpretability to identify key drivers of churn. This proposed approach exhibits potential to provide accurate predictions as well as salient insights, enabling energy providers to design targeted retention strategies.
Downloads
References
[1] Kumar A., Choudhary A. and Chaudhary A. (2023) Churn Prediction Model of Telecom Industry, in Proc. 2023 5th Int. Conf. on Advances in Computing, Communication Control and Networking (ICAC3N), pp. 1–6. doi:10.1109/ICAC3N60023.2023.10541596.
[2] Haddadi S., Farshidvard A., Silva F., Reis J. and Reis M. (2024) Customer churn prediction in imbalanced datasets with resampling methods: A comparative study, Expert Syst. Appl., 246, 123086. doi:10.1016/j.eswa.2023.123086.
[3] Raj A. and Vetrithangam D. (2023) Machine learning and deep learning technique used in customer churn prediction: A review, 2023 Int. Conf. on Computational Intelligence and Sustainable Engineering Solutions (CISES), pp. 139–144. doi:10.1109/CISES58720.2023.101835306.
[4] Srinivasan R., Rajeswari D. and Elangovan G. (2023) Customer Churn Prediction Using Machine Learning Approaches, 2023 Int. Conf. on Artificial Intelligence and Knowledge Discovery in Concurrent Engineering (ICECONF), pp. 1–6. doi:10.1109/ICECONF57129.2023.10083813.
[5] AlShourbaji I., Helian N., Sun Y., Hussien A., Abualigah L. and Elnaim B. (2023) An efficient churn prediction model using gradient boosting machine and metaheuristic optimization, Sci. Rep., 13. doi:10.1038/s41598-023-41093-6.
[6] Hu X., Zhu M., Feng Z. and Stanković L. (2024) Manifold-based Shapley explanations for high dimensional correlated features, Neural Netw., 180, 106634. doi:10.1016/j.neunet.2024.1066344.
[7] Shin J. (2023) Feasibility of local interpretable modelagnostic explanations (LIME) algorithm as an effective and interpretable feature selection method: comparative fNIRS study, Biomed. Eng. Lett., 13, 689–703. doi:10.1007/s13534-023-00291-x.
[8] Bitto A. K., Karim R., Begum M. H., Khan M. F. I. K., Hassan M. M. and Masum A. K. M. (2025) Explainable AI Based Deep Ensemble Convolutional Learning for Multi-Categorical Ocular Disease Prediction, EAI Endorsed Transactions on AI and Robotics, 4. doi:10.4108/airo.9234.
[9] Del Ser J., Barredo-Arrieta A., Diaz-Rodriguez N., Herrera F., Saranti A. and Holzinger A. (2024) On Generating Trustworthy Counterfactual Explanations, Information Sciences, 655, 119898. doi:10.1016/j.ins.2023.119898.
[10] Fekri M. N., Grolinger K. and Mir S. (2023) Asynchronous Adaptive Federated Learning for Distributed Load Forecasting with Smart Meter Data, International Journal of Electrical Power & Energy Systems, 150, 109285. doi:10.1016/j.ijepes.2023.109285.
[11] Wagh S., Andhale N., Deshmukh R., Patil M. and Patil P. (2023) Customer churn prediction in telecom sector using machine learning, Results Control Optim., 10, 100173. doi:10.1016/j.rico.2023.100173.
[12] Su C., Peng X., Yang D., Lu R., Huang H. and Zhong W. (2024) A Transferable Ensemble Additive Network for Interpretable Prediction of Key Performance Indicators, IEEE Trans. Instrum. Meas., 73, 1–14. doi:10.1109/TIM.2024.3472806.
[13] Balaji G., Gowtham N., Tarun S., Prajapati G. and Manikandan N. (2024) Customer Churn Prediction Using Machine Learning Algorithms, in Proc. 2024 Int. Conf. on Emerging Research in Computational Science (ICERCS), pp. 1–6. doi:10.1109/ICERCS63125.2024.10895079.
[14] Ye F., Li X., Zhang N. and Xu F. (2024) Prediction of Single-Well Production Rate after Hydraulic Fracturing in Unconventional Gas Reservoirs Based on Ensemble Learning Model, Processes, 12(6), 1194. doi:10.3390/pr12061194.
[15] Delgado-Panadero A., Benitez-Andrades J. and Garcia-Ordas M. (2023) A generalized decision tree ensemble based on the Neural Networks architecture: Distributed Gradient Boosting Forest (DGBF), Appl. Intell., 53, 22991–23003. doi:10.1007/s10489-023- 04735-w.
[16] Vezzoli M., Zogmaister C. and Van den Poel D. (2020) Will they stay or will they go? Predicting customer churn in the energy sector, Appl. Mark. Anal., 6(2), 136–150. doi:10.69554/HEFD7326.
[17] Douzas G. and Bacao F. (2018) Effective data generation for imbalanced learning using conditional generative adversarial networks, Expert Syst. Appl., 91, 464–471. doi:10.1016/j.eswa.2017.09.030.
[18] Alotaibi M. Z. and Haq M. A. (2024) Customer Churn Prediction for Telecommunication Companies using Machine Learning and Ensemble Methods, Eng. Technol. Appl. Sci. Res., 14(3), 14572–14578. doi:10.48084/etasr.7480.
[19] Hossan M. Z., Riipa M. B., Hossain M. A., Dhar S. R., Zaman A. M., Hossain M., Hossen A. and Sozib H. M. (2025) AI-Powered Predictive Analytics for Financial Risk Management in U.S. Markets, EAI Endorsed Transactions on AI and Robotics, 4. doi:10.4108/airo.9532.
[20] Shirazi F. and Mohammadi M. (2019) A big data analytics model for customer churn prediction in the retiree segment, Int. J. Inf. Manage., 48, 238–253. doi:10.1016/J.IJINFOMGT.2018.10.005.
[21] Gongane V. U.,Munot M. and Anuse A. (2024) A survey of explainable AI techniques for detection of fake news and hate speech on social media platforms, J. Comput. Soc. Sci., 7, 587–623. doi:10.1007/s42001-024-00248-9.
[22] Peng K. and Peng Y. (2022) Research on Telecom Customer Churn Prediction Based on GA-XGBoost and SHAP, J. Comput. Commun., 10(11), 107–120. doi:10.4236/jcc.2022.1011008.
[23] Tiwari U., Ashwani S., Tripathy A. and Kumar K. (2024) A Two-Stage Ensemble Approach for Analysis of Optimizing Customer Churn with Lime Interpretability, in Proc. 2024 IEEE Int. Conf. on Computing, Power and Communication Technologies (IC2PCT), vol. 5, pp. 1540– 1545. doi:10.1109/IC2PCT60090.2024.10486713.
[24] Van Zyl C., Ye X. and Naidoo R. (2024) Harnessing Explainable Artificial Intelligence for Feature Selection in Time Series Energy Forecasting: A Comparative Analysis of Grad-CAM and SHAP, Applied Energy, 353 (Part A), 122079. doi:10.1016/j.apenergy.2023.122079.
[25] Masum A. K. M., Bitto A. K., Talukder S. I., Khan M. F. I., Alam M. S. and Uddin K. M. M. (2025) An Explainable AI Based Deep Ensemble Transformer Framework for Gastrointestinal Disease Prediction from Endoscopic Images, EAI Endorsed Transactions on AI and Robotics, 4. doi:10.4108/airo.9795.
[26] Panigrahi B. S., Kanna R. K., Das P. P., Sahoo S. K. and Dutta T. (2024) Machine Learning Based Intelligent Management System for Energy Storage Using Computing Application, EAI Endorsed Transactions on Energy Web, 11. doi:10.4108/ew.6272.
[27] Rakshit S. and Sengupta A. R. (2025) Comparison of Machine Learning and Deep Learning Models Performance in predicting wind energy, EAI Endorsed Transactions on Energy Web, 12. doi:10.4108/ew.7114.
[28] Afzal M., Rahman S., Singh D. and Imran A. (2024) Cross-Sector Application of Machine Learning in Telecommunications: Enhancing Customer Retention Through Comparative Analysis of Ensemble Methods, IEEE Access, 12, 115256–115267. doi:10.1109/ACCESS.2024.3445281.
[29] Masimov E. (2023) PowerCo Dataset, Kaggle.
[30] Zhang Y., Hong Z. and Chen Z. (2025) Incentives or Time-of-Use Pricing: Strategic Responses to Electricity Demand Response Programs for Energy-Intensive Manufacturers, International Journal of Production Economics, 284, 109588. doi:10.1016/j.ijpe.2025.109588.
[31] Saraji M. K. and Streimikiene D. (2023) Challenges to the Low Carbon Energy Transition: A Systematic Literature Review and Research Agenda, Energy Strategy Reviews, 49, 101163. doi:10.1016/j.esr.2023.101163.
[32] Pulkundwar P., Rudani K., Rane O., Shah C. and Virnodkar S. S. (2023) A comparison of machine learning algorithms for customer churn prediction, in Proc. 2023 6th Int. Conf. on Advances in Science and Technology (ICAST), pp. 437–442. doi:10.1109/ICAST59062.2023.10455051.
[33] Manzoor A., Qureshi M. A., Kidney E. and Longo L. (2024) A Review on Machine Learning Methods for Customer Churn Prediction and Recommendations for Business Practitioners, IEEE Access, 12, 70434–70463. doi:10.1109/ACCESS.2024.3402092.
[34] Alfayan M. and Hagras H. (2025) Towards an Explainable Artificial Intelligence Approach for Smart Grid Systems, Discover Artificial Intelligence, 5, 40. doi:10.1007/s44163-025-00261-5.
[35] Khoh W. H., Pang Y. H., Ooi S. Y., Wang L. Y. K. and Poh Q. W. (2023) Predictive Churn Modeling for Sustainable Business in the Telecommunication Industry: Optimized Weighted Ensemble Machine Learning, Sustainability, 15(11), Art. no. 8631. doi:10.3390/su15118631.
[36] Tonati S., Di Vece M., Pellungrini R. and Giannotti F. (2025) Ensemble Counterfactual Explanations for Churn Analysis, in Discovery Science. DS 2024, Lecture Notes in Computer Science, vol. 15244, Springer, Cham. doi:10.1007/978-3-031-78980-9_21.
[37] Peng K., Peng Y. and Li W. (2023) Research on customer churn prediction and model interpretability analysis, PLoS ONE, 18. doi:10.1371/journal.pone.0289724.
[38] Droste J., Fuchs R., Deters H., Klunder J. and Schneider K. (2024) Explainability Requirements for Time Series Forecasts: A Study in the Energy Domain, in Proc. 2024 IEEE 32nd Int. Requirements Engineering Conf. (RE), pp. 229–239. doi:10.1109/RE59067.2024.00030.
[39] Salih A., Raisi-Estabragh Z., Galazzo I., Radeva P., Petersen S., Lekadir K. and Menegaz G. (2023) A Perspective on Explainable Artificial Intelligence Methods: SHAP and LIME, Adv. Intell. Syst., 7. doi:10.1002/aisy.202400304.
[40] Mujawar S., Sayed M. and Chaware A. (2024) Empowering Decision Making in Healthcare – A Comparative Analysis of eXplainable AI Techniques Using SHAP and LIME for Cancer Patients Data, in Proc. 2024 3rd Ed. of IEEE Delhi Section Flagship Conf. (DELCON), pp. 1–5. doi:10.1109/DELCON64804.2024.10867049.
[41] Ahmed S., Kaiser S., Hossain M. and Andersson K. (2025) A Comparative Analysis of LIME and SHAP Interpreters With Explainable ML-Based Diabetes Predictions, IEEE Access, 13, 37370–37388. doi:10.1109/ACCESS.2024.3422319.
[42] Bokka Y., Mohan R. and Naik R. (2024) SHAP-Driven Interpretability of Autism Risk in Pregnancy Using Explainable AI, in Proc. 2024 Int. Conf. on Integrated Intelligence and Communication Systems (ICIICS), pp. 1– 7. doi:10.1109/ICIICS63763.2024.10859415.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Abdullah Al Mamun, Md Fazla Saim Tanoor, Abdul Kadar Muhammad Masum, Touhid Bhuiyan, Md. Maruf Hassan
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
This is an open access article distributed under the terms of the CC BY-NC-SA 4.0, which permits copying, redistributing, remixing, transformation, and building upon the material in any medium so long as the original work is properly cited.
