Pitch Type Prediction

Overview

Pitch type prediction is a complex challenge since batterymates put a lot of effort into making it difficult. Therefore, the purpose of this project is to develop a framework which is capable of predicting baseball pitch types, including 4-Seam Fastball, 2-Seam Fastball, curveball, slider, and changeup. This project is basically composed of four parts as follows:

1. Model building (LightGBM) and parameter tuning (Bayesian optimization)
2. Imbalance problem handling (SMOTE and class-weight setting)
3. Model evaluation (overall accuracy and confusion matrix)
4. Ensemble learning (soft voting classifier)

Link to the Project

1. https://nbviewer.jupyter.org/github/KUANCHENGFU/Pitch-Type-Prediction/blob/main/Pitch_type_prediction.ipynb