Intro to Predictive Coding: Overview & Interpretation of Terminology June 2014 | Page 12

they can be word combinations, or mathematical values related to groups of words. The computer learns which features are related to documents in each category, and which distinguish between the categories. When a new document is presented for classification, the computer compares the features of that document with the features known to distinguish the categories and then assigns the new document to the appropriate category based on its features.