Taxonomic Interest Classification

Overview

Map a text corpus against a hierarchical interest taxonomy to measure interest diversity (orthogonality) and classify profiles as polymathic or specialist. The core insight: interest diversity is a measurable property of a content distribution, not a subjective impression.

When to Use

• Corpus of user-generated content needs interest profiling
• Need to quantify how diverse or concentrated someone's activity is
• Mapping posts, comments, articles, or other text to interest categories
• Determining polymathic vs. specialist profile classification
• Feeding interest distribution into downstream archetype assignment

Not for:

• Single-document topic extraction (use NMF or LDA instead)
• Keyword frequency analysis without categorical framing
• Content recommendation (this measures what IS, not what SHOULD BE)