Research on CBA League Championship Based on LSTM Neural Network
Download as PDF
Chenkai Ma, Guozhang Zhao, Wenxian Feng
With the continuous development of the CBA league, the demand for team level analysis and evaluation is also increasing. This paper aims at the level of each team's history, combined with the actual situation of the domestic CBA league, gives the mathematical model and programming realization, making the calculation of the probability of winning, predicting the team's ranking, analyzing the team's level and providing reasonable suggestions for the team. may. This paper first uses the LSTM neural network to predict the score of each team's next game, and then uses the principal component analysis method to comprehensively consider the prediction score, the standard deviation and the mean of each team's historical data. Sort the 14 teams by these three indicators. Finally, a comprehensive evaluation of 14 team levels was obtained.
Monte Carlo tree, computer simulation, system clustering, LSTM neural network, principal component analysis