Predict NBA players' career performance based on SVM

摘要

Predicting the National Basketball Association (NBA) players' career performance learning from history data can be amazing for the one's who paying attention to NBA. To predict NBA players' performance, this paper first adopts a new metric to evaluate players' performance based on factorial analysis through combining some popular existing metrics in NBA analysis. Then after defining feature vector according to the new metric to evaluate players' performance, we construct the training set for the player predicted after clustering players. Finally, learning algorithm is implemented based on Support Vector Machine (SVM) to get the prediction of player. The experiments show that this new way is efficient and well-performed.