FubaKI is a Bundesliga prediction program. It allows anyone to predict the results of upcoming matches with machine learning algorithms. FubaKI allows you to first choose the algorithm you want to use, then choose the timeframe you think is most relevant for the algorithm to train with, and as a last step chose the matchup you want to predict. After these 3 simple steps, you have a predicted a winner for that matchup !
Here is an example of our program:
The program allows you to pick between different algorithms to predict the outcome of the chosen game. You can choose between the Baseline algorithm, Poisson Regression Algorithm and the Dixon Coles Algorithm. Because each of those algorithms has different properties, the chosen Timeframe for the crawled data can greatly impact the outcome. Because of this reason, here is a suggested timeframe for each of the Algorithms.
Baseline Algorithm: Previous 8 Days
Poisson Regression: Previous 2 Seasons
Dixon Coles: Largest Possible timeframe
FubaKI supports Python 3.8+.
Installation requires numpy, matplotlib, aiohttp, asyncio, statsmodels, pandas and pyQT5.
-
Clone this repository to your device
-
Install the needed packages via
pip install -e ., this will install all the needed packages. -
Execute the
main.pyfile via the commandteamproject.
The Data that is used by our Algorithms come from the very large source of https://www.openligadb.de/. Openligadb is a community based Database where anyone with an account can add matches and add to the database.