SNOMED CT Entity Linking Benchmark

A benchmark for linking text in medical notes to entities in SNOMED Clinical Taxonomy. #health

Benchmark
Open
180 joined

Model: Section-Aware Super Dictionary Linker
Published

Abstract

A precision-calibrated, section-aware dictionary entity linker for SNOMED CT. It combines SNOMED CT RF2 synonyms (FSN + synonyms across clinical findings, procedures, and body structures) with mention-concept statistics mined from annotated clinical notes. Matching is token-based longest-match with case-sensitive abbreviation handling, section-header exclusion, and dynamic precision-based blacklisting of overly generic terms. Concepts are validated against the current SNOMED CT International Edition for active status.

Submissions (2)

Submissions
Name Macro char IoUSupport-weighted char IoU
Section-Aware Super Dictionary v2
4w 2d ago
0.4179 0.5806
Section-Aware Super Dictionary v1
4w 2d ago
0.4214 0.5831