NHL Power Play Data Analysis & Prediction
Python, BorutaPy, scikit-feature, Fisher Score, ReliefF, Trace Ratio, scikit-rebate
A data analysis of the 2021-2022 NHL season, predicting powerplay goals for upcoming weeks through
ensemble feature selection. Emphasis on BorutaPy. Project includes validation script for baseline
comparison, prediction csv.
The analysis examines 50 features and outputs the selected features into a csv. A custom week difficulty metric is also created and applied as a feature.
See GitHub for detailed description and up to date data.