330.43
004.85
This article addresses the challenge of predicting user churn in computer games using machine learning methods. The analysis of player behavior, taking into account the level of user experience, allowed for the identification of homogeneous audience clusters and the adaptation of modeling approaches for each group. State-of-the-art ensemble learning algorithms, including XGBoost, LightGBM, and CatBoost, were employed alongside class balancing techniques and feature selection methods. Special emphasis was placed on the interpretability of the results: a suite of explainable models was applied, making it possible to identify key risk factors and enhance the transparency of decision-making. The developed models were tested on real data and integrated into the company’s business processes, enabling the timely detection of up to 80% of users prone to churn and contributing to the optimization of retention strategies.
user churn prediction, machine learning, player behavior analysis, explainable AI, ensemble algorithms
1. Arik K., Gezer M., Tayali S.T. The study of indicators affecting customer churn in MMORPG games // Upravlenec. – 2022. – T. 13, № 6. – S. 70–85
2. Bertens P., Guitart A., Periáñez Á. Games and Big Data: A Scalable Multi-Dimensional Churn Prediction Model // Proceedings of 2017 IEEE Conference on Computational Intelligence and Games (CIG). – New York, NY: IEEE, 2017. – P. 33–36
3. Hadiji F., Sifa R., Drachen A. Predicting player churn in the wild // 2014 IEEE Conference on Computational Intelligence and Games (CIG). – Piscataway, NJ: IEEE, 2014. – P. 1–8
4. Kawale J., Pal A., Srivastava J. Churn Prediction in MMORPGs: A Social Influence Based Approach // Proceedings of 2009 International Conference on Computational Science and Engineering (CSE). – Vancouver, BC: IEEE, 2009. – P. 423–428
5. Lee E., Kim B. Profit Optimizing Churn Prediction for Long-Term Loyal Customers in Online Games // IEEE Transactions on Games. – 2020. – Vol. 12, No. 1. – P. 41–53. – DOI:https://doi.org/10.1109/TG.2018.2871215.
6. Lee M., Park S., Lee S. et al. AI-based early detection to prevent user churn in MMORPG // The Korean Journal of Applied Statistics. — 2024. — Vol. 37, No. 4. — P. 525
7. Varun P.N., Yaswanth V., Tapasvi Y.T., Kulkarni V., Padmasree N. Banking Churn Prediction with Random Forest and Discrete Trees in ML // Indiana Journal of Multidisciplinary Research. — 2024. — Vol. 4, No. 3
