Goal and Non-goal Impact Oriented Machine Learning Based Soccer Player Performance Estimation
Keywords:
Machine learning, Soccer team selection, Feature selection, Goal based impact ratings, Non-goal based impact ratings, Estimation of player performanceAbstract
Background: Many machine learning and data mining methods are available to estimate the teams and players’ performance at various levels by using the historical data.
Purpose: Estimating the futuristic performance of the soccer players with adequate accuracy can aid taking critical decisions while selecting an initial team line-up for the upcoming matches.
Methodology: In the proposed work, feature selection has been done conceptually and purely knowledge-driven by the experts to categorize the features into two groups: Non-goal impact and Goal impact. Subsequently, conjunction of the output of four eminent machine learning based feature selection techniques i.e., Lasso, Mutual information, random forest, and RFECV, has been performed to obtain common features. These features have been presented to three machine learning models, i.e., Random forest, Gradient boost, and Linear regression, for three positional player categories, i.e., defenders, midfielders, and attacking forwards.
Results: The best-performing model, i.e., random forest, with an accuracy of 93.45%, has been used to estimate two types of player performance ratings i.e. non-goal based impact ratings (NGI ratings) and goal based impact ratings (GI ratings) to select the team line-up for the upcoming game.
Conclusion: Using mere machine learning for estimating the soccer player performance is not novel, as it may result in goal-biased or position-biased ratings. Since most of the previous work focused on goal oriented features, where defenders or midfielder would appear far weaker contributors in comparison to attacking players. The proposed work also gave due importance to the features having non-goal oriented impact on the game along with traditional goal-oriented impact features.
References
[1] Madan, K., Taneja, K., Taneja, H., Intelligent Decision Making Technique for Handling Game Tactics in Soccer Sports. International Journal of Mathematical, Engineering and Management Sciences,10(5), 1604-1620 (2025). DOI:10.33889/IJMEMS.2025.10.5.076
[2] Rico-González, M., Pino-Ortega, J., Méndez, A., Clemente, F., Baca, A., Machine learning application in soccer: a systematic review. Biology of sport,40(1), 249-263 (2023). DOI: https://doi.org/10.5114/biolsport.2023.112970
[3] Madan, K., Taneja, K., Taneja, H., Intelligent Strategic Planning Method based Algorithm (ISPMA) for Estimation of Soccer Sports Match Outcome. International Journal of Mathematical, Engineering and Management Sciences,10(1), 249-264 (2025). DOI:10.33889/IJMEMS.2025.10.1.014
[4] Arndt, C., Brefeld, U., Predicting the future performance of soccer players. Statistical Analysis and Data Mining: The ASA Data Science Journal, 9(5), 373-382 (2016).
https://doi.org/10.1002/sam.11321
[5] Malikov, D., Kim, J., Beyond xG: A Dual Prediction Model for Analyzing Player Performance Through Expected and Actual Goals in European Soccer Leagues. Applied Sciences, 14(22), 10390 (2024). https://doi.org/10.3390/app142210390
[6] Kulakou, S., Ragab, N., Midoglu, C., Boeker, M., Johansen, D., Riegler, M. A., Halvorsen, P., Exploration of different time series models for soccer athlete performance prediction. Engineering Proceedings,18(1), 37 (2022). https://doi.org/10.3390/engproc2022018037
[7] Wiik, T., Johansen, H. D., Pettersen, S. A., Baptista, I., Kupka, T., Johansen, D., Halvorsen, P., Predicting peek readiness-to-train of soccer players using long short-term memory recurrent neural networks. In 2019 International Conference on Content-Based Multimedia Indexing (CBMI)(pp. 1-6). (2019) IEEE.
[8] Bergkamp, T. L., Frencken, W. G., Niessen, A. S. M., Meijer, R. R., den Hartigh, R. J., How soccer scouts identify talented players. European Journal of Sport Science,22(7), 994-1004 (2022). https://doi.org/10.1080/17461391.2021.1916081
[9] Dijkhuis, T. B., Kempe, M., Lemmink, K. A., Early prediction of physical performance in elite soccer matches—a machine learning approach to support substitutions.Entropy,23(8), 952 (2021). https://doi.org/10.3390/e23080952
[10] Carpita, M., Ciavolino, E., Pasca, P., Players’ role-based performance composite indicators of soccer teams: A statistical perspective. Social Indicators Research,156(2), 815-830 (2021). https://doi.org/10.1007/s11205-020-02323-w
[11] Dick, U., Tavakol, M., Brefeld, U., Rating player actions in soccer. Frontiers in Sports and Active Living,3, 682986 (2021). https://doi.org/10.3389/fspor.2021.682986
[12] Stafylidis, A., Mandroukas, A., Michailidis, Y., Vardakis, L., Metaxas, I., Kyranoudis, A. E., Metaxas, T. I., Key performance indicators predictive of success in soccer: A comprehensive analysis of the Greek soccer league. Journal of Functional Morphology and Kinesiology,9(2), 107 (2024). https://doi.org/10.3390/jfmk9020107
[13] Pantzalis, V. C., Tjortjis, C., Sports analytics for football league table and player performance prediction. In 2020 11th International Conference on Information, Intelligence, Systems and Applications (IISA(pp. 1-8) (2020). IEEE. DOI: 10.1109/IISA50023.2020.9284352
[14] Nogueira, S., Sechidis, K., Brown, G., On the stability of feature selection algorithms. Journal of Machine Learning Research,18(174), 1-54 (2018). https://www.jmlr.org/papers/v18/17-514.html
Downloads
Published
License
Copyright (c) 2026 Karun Madan, Dr. Kavita Taneja, Dr. Harmunish Taneja
This work is licensed under a Creative Commons Attribution 4.0 International License.
Articles in the Graduate Journal of Interdisciplinary Research, Reports and Reviews (Grad. J. InteR3) by Vyom Hans Publications are published and licensed under a Creative Commons Attribution- CC-BY 4.0 International License.
