This python module implements a non-optimized version of the ID3 Algorithm to create a binary decision tree. Rather than using a traditional entropy measure, this algorithm uses Pearson's Chi-Squared relevancy test. If a feature meets the user-defined p-value, a new decision node is created in the classifier tree.

The sample dataset is from the KDD Cup 2000, and gives the clickstream and purchase behavior of users on Gazelle.com. In testing, the ID3 predictor estimated user behavior with a roughly 75% accuracy.