In this repository, I implemented the PitchNet, a pitch classification neural network mentioned in this article written by a data scientist from the MLB. The PitchNet is implemented using the Keras functional API and achieves a 95% accuracy on a 9-class classfication task.
The data is scraped from the baseball website Savant, and includes all pitch-by-pitch data from 2014 to 2019.
The neural network takes three continuous input: the release speed, horizontal movement, and vertical movement of a pitch, and one categorical input: the pitcher's ID, which is fed into an embedding layer.
You can view the notebook here.