CBA Games Result Prediction Based on One-dimensional Convolutional Neural Network
As a competitive sport,results of the basketball game is so important to the team and the spectators,and has such an impact on the sports betting,so that predicting the results of the game becomes one of the main focuses of attention outside of the event.Machine learning and deep learning are used to predict the results of the game,this paper proposes a method for pre-dicting the results of CBA games based on one-dimensional convolutional neural network,after the Pearson correlation coeffi-cient calculation and principle component analysis(PCA)dimensionality reduction,the regular season data of 20 teams in the past 5 seasons of the CBA are input into the self-built ST-1D-CNN model for training to achieve the prediction of the results of the games,and the model reaches an accuracy rate of 76.50%.
basketballone-dimensional convolutional neural networkPearson correlation coefficientprinciple component analysisprediction of game result