Create
Design Project Gallery
Predicting Pitcher Injury: A survival analysis approach
- Program: Computer Science
- Course: Other
- Year: 2025
Project Description:
Baseball pitchers, at all ability levels, are injuring their arms at an alarming rate. Thus predicting if a pitcher is likely to become injured in the near future would be helpful to the pitcher’s career and health. Traditional statistical and machine learning models have made strides in injury forecasting; yet they fail to account for real-world complexities, such as censored data, recurring injuries, and the dynamic nature of player workloads. In this work, we apply survival analysis to the problem, benchmarking traditional statistical approaches with modern recurrent neural networks (RNNs). We apply these models to Statcast data collected from professionals leagues, demonstrating the superiority of survival RNNs.
Project Photo:
Pitcher Throwing Baseball with Extreme High Elbow Torque
