SYSTEM AND METHOD FOR CATEGORY DISCOVERY USING STATISTICAL ANALYSIS FOR TERM COOCCURRENCE

摘要

A system and method for category discovery is disclosed. The method (600) discloses: receiving (602) an information collections including a set of strings: identifying (604) positively predictive pairs of strings; identifying (606) negatively predictive pairs of strings; joining (608) positively predictive pairs of strings into a category; and splitting (610) negatively predictive pairs of strings into different categories. The system (100) discloses various elements, means and instructions for performing the method.