These are the supporting files used for data processing and training machine learning models for this CS221 project on detecting cheating in multiplayer online games.
- Contains the player beahvior collection server.
- To run, Node.js and NPM must be installed. You can
npm install && npm start
inside theapi
directory to start the server.
- Contains data files that were in a format usable by R to run the stray algorithm, which can be installed following the instructions on their GitHub repostiroy.
- Sample commands used to find and visualize anomalies can be found in
stray.r
.
- Contains raw game state data and Python scripts used to preprocess the data and train an SVM classifier in the
data
directory.- Kills are stores in separate CSVS than game state data.
- The
preprocess.py
script will compute game state gradients - The
svm.py
script will train an SVM using SKLearn on the training data output from preprocessing.