Code requires scikit library and python 2.7
*Scrape weekly data with scripts in ruby_code
Quick summary: Simulate.py has the code for specifying the simulation,prints number of games won by each team Game simulation is in game.py Agent=Team class in team.py data_ops.py handles the parsing and retrieving of training data learn.py has the methods for completion probabilty and training a KNN model. plays.py has method of executing a play, although this is only used by the environment, team.py has play methods for agents
Testing: regression_testing.py has the framework that was used to test rushing regression pass_regression_testing.p has the same but for passing knn.py has the code used for testing KNN algorithms and SVMs SVM_testing has the orginal code for testing SVMs hmm.py has some messing around with HMM's
Data files. dstats.py is used for loading defensive team data. data/ directory stores the parsed training data and other data files data_scripts/ has the files used for scraping data and parsing
The simulation code is in game.py and team.py. Each team/agent is part of the Team class in team.py, and this is where the decision by an agent to choose a play is made.