Neural Network: Creates a Convolutional Neural Network to predict outcomes.

Decision Tree: Works by splitting data over specific parameters i.e. all teams that shoot better than 40% go one way, the rest go the other way, repeating this process until fitted.

Linear Regression: Implements Linear Regression for regression and RidgeClassificationCV for Classification, mapping each parameter to a coefficient linearly.

K Nearest: Maps each point in n-dimensional space and returns the average/mode of the k-nearest other vectors.

XG Boosted Tree: Essentially a fancy Decision Tree which performs better by running gradients on successive trees.

Random Forest: A collection of decision trees, takes longer to train but is more accurate. (Also gives probability values)

Pros and Cons: XG Boosted and Linear Regression give good results for minimal training time, while Random Forest can give better results (2-3% better) for longer training times

Score: This method does regression to predict scores of both teams.

Winner: This method uses binary classification to predict the winner.

Pros and Cons: The Winner method performs on average 6% better (73.5% to 66.7%) compared to Score

Welcome to the Parameter Selection!

You'll be selecting options for the Prediction Method, Model Type, and Parameters.

• Prediction Method: Choose between predicting each team's score or simply which team wins.
• Model Type: Select the type of model used for predictions.
• Parameters: Configure parameters specific to the selected model.
• Number of Models: Train multiple models and pick the best one.

If unsure, using linear regression to predict the winner is a solid choice!