Using CFBFastR and, I chose from a variety of categories. When selecting predictors I wanted to make sure that I picked out stats that accurately describe how well a team played in various facets of the game and give context to the talent on each team.
In this article I will describe the method and data that the model uses to make predictions, go into more detail about the model’s accuracy, and then give the model’s picks for the 2021 season. Using an Extreme Gradient Boosting (XGBoost) Algorithm, I predicted college football win totals based on various team statistics, recruiting, and rankings data.
