MAKİNE ÖĞRENMESİ YÖNTEMLERİYLE EUROLEAGUE BASKETBOL MAÇ SONUÇLARININ TAHMİN EDİLMESİ VE MAÇ SONUÇLARI ÜZERİNDE EN ETKİLİ DEĞİŞKENLERİN BULUNMASI

Abstract

Bu çalışmada 2016-2017 ile 2020-2021 yılları arasında oynanan 1358 EuroLeague basketbol maçlarındaki takım istatistikleri göz önüne alınmış ve bu takım istatistiklerinden hangilerinin maçın galibi üzerinde en çok etkiye sahip olduğu belirlenmeye çalışılmıştır. Maçlar, k-ortalama kümeleme analizi sonucunun belirttiği skor farklarına göre yakın, dengeli ve dengeli olmayan olmak üzere üç gruba ayrılmıştır. Hem bu üç grup hem de tüm maçlara k en yakın komşuluk, naive bayes, lojistik regresyon, destek vektör makinaları, karar ağacı, rastgele orman ve yapay sinir ağları algoritmaları uygulanmış ve en etkili algoritmalar lojistik regresyon, destek vektör makineleri ve yapay sinir ağları olarak bulunmuştur. Bu üç algoritma maç sonucunu tüm maçlar için yaklaşık %84 oranında doğru bilmiştir. Yakın maçlarda bu oran %79 a düşmüş, dengeli maçlarda %97 e, dengeli olmayan maçlarda %100 e çıkmıştır. Maç sonucu üzerinde en çok etkili olan değişkenler savunma ribaundu, gerçek şut yüzdesi, top çalma, top kaybı, hücum ribaundu ve denenen serbest atıştır. Burada bulunan sonuçlar takımların maç içi stratejilerini belirlerken en çok odaklanmaları gereken konular üzerinde fikir vermekte ve bu konular üzerine yoğunlaşarak strateji belirlemelerine yardımcı olacağı düşünülmektedir.

Keywords

• Basketbol
• makine öğrenmesi
• maç sonucu tahmini
• Euroleague

PREDICTING EUROLEAGUE BASKETBALL MATCH OUTCOMES WITH MACHINE LEARNING TECHNIQUES AND REVEALING THE MOST IMPORTANT GAME RELATED VARIABLES

Abstract

In this study, team statistics in 1358 EuroLeague basketball matches played between 2016-2017 and 2020-2021 seasons were taken into account and it was tried to determine which of these team statistics had the most impact on the winner of the match. The matches were divided into three groups as close, balanced and unbalanced games according to the score differences indicated by the k-means cluster analysis result. K nearest neighbor, naive bayes, logistic regression, support vector machines, decision tree, random forest and artificial neural network algorithms were applied to both these three groups and all matches, and the most effective algorithms were found to be logistic regression, support vector machines and artificial neural networks. These three algorithms can predict correctly the match result for all matches with approximately 84% accuracy. This rate decreased to 79% in close matches, increased to 97% in balanced matches and to 100% in unbalanced matches. The variables that have the most influence on the outcome of the match are defensive rebounds, true shooting percentage, steals, turnovers, offensive rebounds and free throw attempts. The results give an idea on the issues that the teams should focus on while determining their in-match strategies and help them determine their strategy by focusing on these issues.

Keywords

• Basketball
• machine learning
• predict match outcome
• Euroleague

References

1. Bishop C.M. (2006). Pattern Recognition and Machine Learning, Springer.
2. Çene E. (2018). What is the difference between a winning and a losing team: insights from Euroleague basketball. International Journal of Performance Analysis in Sport, 18(1). doi:10.1080/24748668.2018.1446234
3. Csataljay G, O’Donoghue P, Hughes M, ve Dancs H. (2009). Performance indicators that distinguish winning and losing teams in basketball. International Journal of Performance Analysis in Sport, 9(1), 60–66. doi:10.1080/24748668.2009.11868464
4. Davoodi E ve Khanteymoori AR. (2010). Horse racing prediction using Artificial Neural Networks. Proc. of the 11th WSEAS Int. Conf. on Neural Networks, NN ’10, Proceedings of the 11th WSEAS Int. Conf. on Evolutionary Computing, EC ’10, Proc. of the 11th WSEAS Int. Conf. on Fuzzy Systems, FS ’10 içinde (ss. 155–160).
5. Garcia J, Ibanez SJ, Gomez MA, ve Sampaio J. (2014). Basketball Game-related statistics discriminating ACB league teams according to game location, game outcome and final score differences. International Journal of Performance Analysis in Sport, 14(2), 443–452. doi:10.1080/24748668.2014.11868733
6. Gevrey M, Dimopoulos I, ve Lek S. (2003). Review and comparison of methods to study the contribution of variables in artificial neural network models. Ecological Modelling içinde (C. 160, ss. 249–264). Elsevier. doi:10.1016/S0304-3800(02)00257-0
7. Gorunescu F. (2011). Classification performance evaluation. Data Mining. Intelligent Systems Reference Library, 12(319–330). Springer, Berlin, Heidelberg.
8. Hastie T, Tibshirani R, James G, & Witten D. (2021). An Introduction to Statistical Learning (2nd Edition). Springer Texts. New York.

9. Horvat T, Havaš L, ve Srpak D. (2020). The impact of selecting a validation method in machine learning on predicting basketball game outcomes. Symmetry, 12(3), 431. doi:10.3390/sym12030431
10. Ibáñez S, Sampaio J, Feu S, Lorenzo A, Gomez M, ve Ortega E. (2008). Basketball game-related statistics that discriminate between teams’ season-long success. European Journal of Sport Science, 8(6), 369–372. doi:10.1080/17461390802261470
11. Jones ES. (2016). Predicting Outcomes Of NBA Basketball Games. North Dakota State University.
12. Kaur H ve Jain S. (2018). Machine learning approaches to predict basketball game outcome. Proceedings - 2017 3rd International Conference on Advances in Computing, Communication and Automation (Fall), ICACCA 2017, 2018-Janua, 1–7. doi:10.1109/ICACCAF.2017.8344688
13. Kuhn M. (2008). Building predictive models in R using the caret package. Journal of Statistical Software, 28(5), 1–26. doi:10.18637/jss.v028.i05
14. Leicht AS, Gómez MA, ve Woods CT. (2017). Explaining Match Outcome during The Men’s Basketball Tournament at The Olympic Games. Journal of Sports Science and Medicine, 16(August), 468–473.
15. Loeffelholz B, Bednar E, ve Bauer KW. (2009). Predicting NBA Games Using Neural Networks. Journal of Quantitative Analysis in Sports, 5(1). doi:10.2202/1559-0410.1156
16. Lorenzo A, Gómez MÁ, Ortega E, Ibáñez SJ, ve Sampaio J. (2010). Game related statistics which discriminate between winning and losing under-16 male basketball games. Journal of Sports Science and Medicine, 9(4), 664–668.
17. Magel R ve Unruh S. (2013). Determining Factors Influencing the Outcome of College Basketball Games. Open Journal of Statistics, 03(04), 225–230. doi:10.4236/ojs.2013.34026
18. McCabe A ve Trevathan J. (2008). Artificial intelligence in sports prediction. Proceedings - International Conference on Information Technology: New Generations, ITNG 2008 içinde (ss. 1194–1197). doi:10.1109/ITNG.2008.203
19. McComb DG. (2004). Sports in world history. Sports in World History. Routledge Taylor & Francis Group. doi:10.4324/9780203697016
20. Navega D, Coelho C, Vicente R, Ferreira M.T., Wasterlain S, & Cunha E. (2015). AncesTrees: ancestry estimation with randomized decision trees. International Journal of Legal Medicine, 129(5), 1145–1153.
21. Oliver D. (2004). Basketball on Paper: Rules and Tools for Performance Analysis. Potomac Books, Inc..
22. Ozkan IA. (2020). A Novel Basketball Result Prediction Model Using a Concurrent Neuro-Fuzzy System. Applied Artificial Intelligence, 34(13), 1038–1054. doi:10.1080/08839514.2020.1804229
23. Pai PF, ChangLiao LH, ve Lin KP. (2017). Analyzing basketball games by a support vector machines with decision tree model. Neural Computing and Applications, 28(12), 4159–4167. doi:10.1007/s00521-016-2321-9
24. Rasouliyan L & Miller DP. (2006). The logic and logistics of logistic regression. Ovation Research Group, San Fransisco California, 1–14.
25. Smola A.J. & Schölkopf B. (2004). A tutorial on support vector regression. Statistics and Computing, 14(3), 199–222.
26. Tax N ve Joustra Y. (2015). Predicting The Dutch Football Competition Using Public Data: A Machine Learning Approach. Transactions on knowledge and data engineering, 10(10), 1–13. doi:10.13140/RG.2.1.1383.4729
27. Thabtah F, Zhang L, ve Abdelhamid N. (2019). NBA Game Result Prediction Using Feature Analysis and Machine Learning. Annals of Data Science, 6(1), 103–116. doi:10.1007/s40745-018-00189-x
28. Valenzuela R. (2018). Predicting national basketball association game outcomes using ensemble learning Techniques, Yayımlanmamış Yüksek Lisans Tezi, California State University.
29. Wiseman O. (2016). Using Machine Learning to Predict the Winning Score of Professional Golf Events on the PGA Tour.