Taxonomies and Ontologies
Part of the Query Understanding series by Daniel Tunkelang.
Overview
Taxonomies and ontologies are two types of structured knowledge representation that help search systems understand the relationships between concepts. A taxonomy organizes concepts into a strict hierarchy — a tree of categories where each concept belongs to one parent — and is the basis for catalog navigation and for scoping queries to the right part of a collection. An ontology is richer: it captures not just hierarchical relationships but also equivalences, part-whole relationships, compatibility constraints, and other semantic connections between concepts. Both allow the system to reason about what a user probably means beyond the literal words they typed — inferring related categories, finding synonymous terms, or understanding that one thing is a subtype of another. Maintaining these structures as a catalog evolves is an ongoing operational challenge.
Note: Article content behind Medium paywall — accessible at source URL with Medium account.