NBA_Predictions | Reworked NBA Predictions

NBA_Predictions is a Python library. NBA_Predictions has no bugs, it has no vulnerabilities and it has low support. However NBA_Predictions build file is not available. You can download it from GitHub.

The idea of this project was to create a naive nba predictor. There is very little knowledge about basketball programmed into this algorithm. For example, many algorithms assume feature importance (E.g. Three point % is more important than the winning streak…​), this is not the case for this algorithm. No feature importances are given, and it is left up to the algorithm to find the best fit. Here is a summary of the statistics used: - Player Statistics: See: for the regular and advanced statistics. - Team Statistics: contains the regular, per game, and advanced statistics. - Schedule Statistics: These are only three statistics: (1) Home/Away, (2) Win/Lose streak, and (3) # days off prior to game. The player statistics are modified by the current NBA injury reserve list. Here, some injuries are worth more than others, as a rolled ankle is significantly less impactful than a play-preventing injury. A genetic algorithm is used on a matrix function of the above statistics. The coefficients evolve to predict the prior games as best as possible (weighting recent games higher). After a set number of generations, the algorithm predicts out the next week. Then for the current day, the online sports book odds are downloaded, compared and the best games to bet on are outputed for the day.