A custom bayesian network model for predictions.
Predicts past and future matches based on the data parsed from the official overwatch league api (api.overwatchleague.com/).
Prediction Accuracy on OWL 2019 Stage 2 Games:
- Average Win/Lose Accuracy - 74%
- Average Match Score Accuracy - 42%
Requirement:
Python: Python 2.x
Packages: Numpy, Pillow
pip install -r requirements.txt
Usage:
python predict.py
Options:
-c --collect Collect new data from OWL api
-l --list Display a list of team names and abbrs
-m --manual Run a single match prediction with maunal inputs
-n --new Batch run match predictions on a future week's schedule
-s --stage Specify Stage
-w --week Specify Week
-e --eval Evaluate output results
-i --image Enable image output
Sample Output:
-
Text:
-
Image:
Bayesian Network Model:
