RelAgent: Entity Detection and Normalization for Diseases in Clinical Records: a Linguistically Driven Approach
SV Ramanan, Senthil Nathan · 2014
We refined the performance of Co-coa/Peaberry, a linguistically moti-vated system, on extracting disease en-tities from clinical notes in the train-ing and development sets for Task 7. Entities were identified in noun chunks by use of dictionaries, and events (‘The left atrium is dilated’) through our own parser and predicate-argument struc-tures. We also developed a mod-ule to map the extracted entities to the SNOMED subset of UMLS. The module is based on direct matching against UMLS entries through regu-lar expressions derived from a small set of morphological transformations, along with priority rules when multi-ple UMLS entries were matched. The performance on training and develop-ment sets was 81.0 % and 83.3 % respec-tively (Task A), and the UMLS match-ing scores were respectively 75.3 % and 78.2 % (Task B). However, the perfor-mance against the test set was low by comparison, 72.0 % for Task A and 63.9 % for Task B, even while the pure UMLS mapping score was reasonably high (relaxed score in Task B = 91.2%). We speculate that our moderate perfor-mance on the test set derives primarily from chunking/parsing errors. 1