resnik-1995-similarity-formula

IN premise — summaries/2026/08/24/wiki-Semantic_similarity-chunk-1.md

Created 2026-08-25T02:58:52+00:00

Resnik's (1995) similarity measure is defined as sim(A,B) = IC(LCS(A,B)), using only the information content of the lowest common subsumer.

Summary

Two concepts are judged similar or dissimilar based entirely on how specific their shared category is, ignoring everything the concepts themselves are. This means the system's notion of resemblance is anchored to hierarchy: two things that both fall under a narrow, rare category look alike to the system, even if they differ in every other respect, while two things that only share a vague umbrella category like "entity" are treated as dissimilar no matter how alike they otherwise appear.