Another year, another chance to predict the upsets, call the probabilities and put your bracketology skills to the leaderboard test. In this keras deep learning project, we will once again attempt to predict the outcomes of this year's US men's college basketball tournament. But unlike most deep learning projects, we will pick the winners and losers using a combination of rich historical data and computing power, while the ground truth unfolds on national television.
If you are unfamiliar with the format and intricacies of the NCAA tournament, we encourage reading the wikipedia page before diving into the data. The data description and schema may seem daunting at first but is not as complicated as it appears.
As a reminder, you are encouraged to incorporate your own sources of data. We have provided team-level historical data to jump-start the modeling process, but there is also player-level and game-level data that may be useful.
We extend our gratitude to Kenneth Massey for providing much of the historical data.