jcwhuang / cs224w-project

Final project for CS224W at Stanford University, Autumn 2015

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

A network-assisted approach to predicting passing distributions

Final project for CS224W at Stanford University

###Autumn 2015###

Data: UEFA Champions League 2014-15 season. Relevant data can be found in the data directory in the following subfolders:

  • passing distributions: passing_distributions/
  • rankings: rankings/
  • squad lists: squads/
  • scores: scores/
  • tactical lineups: lineup/
  • individual player statistics: player_statistics/

Goal: Based on games from the group stage, predict passing distributions for games in the round of 16 stage

Results: Our linear regression model exceeded our baseline (average past passing network) by 25.27%

To duplicate our results or run experiments of your own with our linear predictor:

Linear predictor with entire model sharing one set of weights

$ cd predicted
$ python pdPrediction.py

Linear predictor with each team having its own set of weights

$ cd predicted
$ python pdPrediction-team.py

For more details including model, algorithm, analysis, results, see writeup/final.pdf

About

Final project for CS224W at Stanford University, Autumn 2015


Languages

Language:Python 82.3%Language:TeX 17.1%Language:Shell 0.6%