README
======

Answering the question "Will my team win?" using logistic regression

Our project is a machine learning module that trained on data from URF games.
Given a set of 5 champions, our service will predict the result of a match, 
i.e. victory or defeat.

We used Python and MATLAB.

System requirements: Python 2

To run, open terminal and cd to the directory containing teamEval.py. 
Run 'python teamEval.py'.

* To reconstruct the training process, all paths to local files used in 
each script must be modifed to fit the appropriate environment.

scraper.py
----------
A Python script used to pull URF game IDs from the endpoint. These IDs are
written to a file on disk. We left this script running for over a week. 
We ended up collecting almost 60,000 game IDs.

fetcher.py
----------
A Python script used to pull game data given a list of game IDs. After URF
concluded, we used this script to pull game data through the Riot API.
We constructed feature vectors for our design matrix as we pulled the game data.
This script also writes the final design matrix and its respective labels
for training into a .mat file.

regression.m
------------
A MATLAB script we used to perform find the beta-vector for our logistic regression
classfier.

teamEval.py
-----------
The main interface of this project. It contains a hard-coded version of the beta
vector found using regression.m. Run our project by using 
python teamEval.py

Team Members:
-----------
Leo Kwan, CA USA
Codelyouko, NA

James Wei, CA USA
AWeiOfLife, NA