This data project utilizes the use of machine learning to predict the type of pitch thrown from a MLB pitcher based on the characteristics of the pitch (e.g., initial speed, horizontal break, vertical break).
The use of machine learning models in baseball have become popular thanks to the efficient 'learning' capabilities of these models to learn what an outcome (e.g., pitch type) should be based on a number of features (e.g., initial speed, horizontal break) fed into the model. This is considered training the model. The goal is to train a model to be as accurate as possible when predicting an outcome. Here I show a basic example of using general and pitcher-specific machine learning models to predict a pitcher's pitch type (i.e., two-seam fastball, four-seam fastball, curveball) based on the characteristics of the pitch. The data consists of six different pitchers.