Summary

Our group used data collected by a basketball fan and a former analyst for the Dodgers to build a model to predict the outcome of any basketball game, using both team and player statistics. The data provides different metrics of both teams playing in each game. After creating two datasets for training and testing, we used different models and classification methods to predict outcomes of each game and compared accuracy rates of each method used. Eventually, the method which gave us the best accuracy was Principal Component Regression with close to 70% accuracy rate.