The features.csv file contains a list of protein binding sites and non-binding sites. The features are represented by 'y' and 'n' which indicate the presence or absence, respectively, of amino acids within various shell distances away from the center of site.
bayes_model.py creates and trains a Bayesian predictor, outputs a confusion matrix to access accuracy. Also is able to sort the data by amino acids and shell space to see which features are most indicative of protein binding site identity.