Project to Predict the outcomes of baseball games using machine learning. You can see the full details on the technical paper here.

In this paper we investigate using machine learning algorithms to predict the outcomes of baseball games. Baseball has a large amount of raw data, including pitching, batting, and defensive statistics for each game. In addition, baseball has the largest total number of games per season of any sport. This combination of readily available data and the large number of games make baseball a great prospect for machine learning. We use past data to predict the outcomes of baseball games with the goal of discerning any patterns or shedding light on what characteristics produce a winning baseball team. We use logistic boost and SVMS

This project was done for the graduate Machine Learning course at Stanford (CS 229, Fall 2011).