PREDICTION OF THE NUMBER OF INDEX POINTS OF PLAYERS IN THE ABA LEAGUE WITH A FOCUS ON DATA COLLECTION AND EXPLORATORY ANALYSIS
DOI:
https://doi.org/10.24867/13BE33NisicKeywords:
Lasso regression, Random Forest regression, LightGBM regressor, Personal index rating predictionsAbstract
This paper presents the machine learning approach to automatically predict the number of index points that a player achieves in a basketball game. The focus of the work is on data collection and exploratory analysis. Data was collected from eurobasket.com using web-scraping techniques. After cleaning the data set, feature extraction and exploratory data analysis were performed. The prediction was made using three different regressors: Lasso, Random Forest, and LightGBM By optimizing the hyperparameters of these algorithms' implementations, we came up with models that were used to predict the number of index points. The LASSO regression model achieved the best results with the mean absolute error MAE = 5.617. The paper proposes further improvements of the data set as future work.
References
[2] Oliver, Dean. Basketball on paper: rules and tools for performance analysis. Potomac Books, Inc., 2004.
[3] Kubatko, Justin, et al. "A starting point for analyzing basketball statistics." Journal of Quantitative Analysis in Sports 3.3 (2007).
[4] Page, Garritt L., and Fernando A. Quintana. "Predictions based on the clustering of heterogeneous functions via shape and subject-specific covariates." Bayesian Analysis 10.2 (2015): 379-410.
[5] South, Charles, et al. "A Starting Point for Navigating the World of Daily Fantasy Basketball." The American Statistician 73.2 (2019): 179-185.
[6] FiveThirtyEight's Career-Arc Regression Model Estimator with Local Optimization, https://projects.fivethirtyeight.com/carmelo/
[7] Casals, Martí, and A. Jose Martinez. "Modelling player performance in basketball through mixed models." International Journal of Performance Analysis in Sport 13.1 (2013): 64-82.
[8] Cai W, Yu D, Wu Z, Du X, Zhou T. A hybrid ensemble learning framework for basketball outcomes prediction. Physica A: Statistical Mechanics and its Applications. 2019 Aug 15;528:121461.
[9] Thabtah F, Zhang L, Abdelhamid N. NBA game result prediction using feature analysis and machine learning. Annals of Data Science. 2019 Mar 7;6(1):103-16.
[10] Shreyas S. Shivakumar. “Learning to Turn Fantasy Basketball Into Real Money.” https://shreyasskandan.github.io/Old_Website/files/report-ChanHuShivakumar.pdf
[11] Porter, Jack W. "Predictive Analytics for Fantasy Football: Predicting Player Performance Across the NFL." (2018).
[12] Lutz, Roman. "Fantasy football prediction." arXiv preprint arXiv:1505.06918 (2015).
[13] Kengo Arao, https://github.com/KengoA/fantasy-basketball/blob/master/report.pdf
[14] Beautiful soup, https://www.crummy.com/software/BeautifulSoup/bs4/doc/
[15] Selenium Web Driver, https://www.selenium.dev/documentation/en/
[16] Scikit-learn, https://scikit-learn.org/
[17] f_regression, https://scikit-learn.org/stable/modules/generated/sklearn.feature_selection.f_regression.html#sklearn.feature_selection.f_regression
[18] Hyperopt, https://github.com/hyperopt/hyperopt
