Using Logic Regression Algorithm to Analyze and Predict NBA Game Data

Abstract

By collecting the basic statistics of NBA players, teams, seasons, and league matches over the last three years, The data were analyzed by K-Means method, The analysis results are sorted according to the strength of the team and divided into five grades; Each team's past matches and Elo grades are used as characteristic vectors, The relationship between each game and the winning team is analyzed by machine learning logic regression method, Finally, through the python programming can predict probability of each team in each game, and with all the strength of the team was compared to verify the prediction results.