Machine Learning Based Prediction OfThe Best Suitable Playing Positions of The Players In The Game of Basketball

The present study highlights the assessment of the best suitable playing positions for the basketball players with the aid of their anthropometric, motor fitness, and Physiological aspects. 250 players from engineering colleges in the Karnataka South Zone Inter-College Basketball tournament were considered for this study. The anthropometric, Physiological, and motor fitness parameters (28 features altogether) were analyzed with the aid of various machine learning models to predict the best suitable playing positions of the players. An 80:20 approach was used to train and test the models, Multilayer perceptron model, linear support vector machine approach, and the Decision tree classifier were the models used. They provided an accuracy of 76%, 82%, and 90% respectively. These results are useful for basketball training protocols and can aid the coaches to deploy their players at respective playing positions to elicit the best performance from the players during the game of basketball.