This study aims to develop and implement a machine learning model to predict the number of wins an NCAA Division I basketball team will achieve in a given season. Utilizing the College Basketball Dataset, we leverage historical data from the 2024 season. The dataset consists of 24 variables relevant to team performance, including valuable variables such as the team’s points scored/allowed per 100 minutes, effective field goal percentage made/allowed, and power rating, representing a team’s probability of defeating an average Division I opponent. The target variable for prediction is the total number of wins in a given season. Given the inherent variability and randomness in sports outcomes, we aim to build a model to predict accurately how many wins a team will get in the season.
