NHL-Prediction-Model icon indicating copy to clipboard operation
NHL-Prediction-Model copied to clipboard

Model for computing the probability of each team winning a given matchup in the National Hockey League

NHL Game Prediction Model

The code here is for creating a model that predicts the outcome of a single NHL game. Using this model we can get the probability that the home and away teams have of winning that matchup. The methodology behind the model can be found here.

Running

All the data used in the analysis is provided in order to verify the conclusions stated in the previous post. The models can be created by running the build_models.py file in the models directory (this will also print the evaluation metrics).

For those looking to run it yourself I recommend looking at the blog post and the code for the variables needed for each game.