Inspiration
Today is the Superbowl, let's see if our network can guess the winner!
What it does
Using our website, you can select two teams and determine the network's best guess for the winner and confidence percentage.
How we built it
Our program takes marginal team statistics from a large data set over many games. These are passed as features to corresponding labels to train the network using TensorFlow. Once sufficiently trained and configured with appropriate hidden layers, batchs and training iterations, the network can take two team names, calculate their cumulative marginal stats and predict the winner.
Challenges we ran into
Formatting inputs in tensorflow with slicing. Passing ID keys by accident that train the network on team hierarchies verses marginal stats of any two teams.
Accomplishments that we're proud of
We got the network to learn!
What we learned
According to our network, the Eagles are more likely to win today
What's next for NFLSenpai?
Keep developing our network and site. Website should be live soon under domain www.nflsenpai.com Maybe we will investigate a NHLGuru? Or even a NBANinja?
Log in or sign up for Devpost to join the conversation.