의료 정보검색에서 효과적인 질의어 확장을 위한 축약된 시소러스의 개발 및 평가
한승빈1), 최명선1), 김재필1), 최진욱*1)
1)서울대학교 의과대학 의공학교실
Abstract : We present a medical document retrieval system, which utilizes the UMLS Metathesaurus, and that incorporates query expansion methodology. To improve the retrieval performance of relevant documents and to reduce retrieval time, we constructed a condensed Metathesaurus (CMT), consisting of practically used terms in medical records and then evaluated its retrieval performance and time for relevant documents. The devised system had average 22% effectiveness for searching relevant documents in the UMLS Metathesaurus (MT) implemented information retrieval (IR) system for query expansion. However, MT implemented system took 3.5 times longer to retrieve documents than a non-MT implemented IR system. When we applied the CMT into IR system, the retrieval time reduced by 50% and the retrieval performance reduced by 8.7% compared with all MT implemented IR system. We expect that the UMLS Metathesaurus and its sophisticated methods for information retrieval will make good use of medical information systems. keyword : Unified Medical Language System, Metathesaurus, Information Storage and Retrieval, Evaluation Studies, Computing Methodologies
투고내용을 삭제하시려면 "삭제"버튼을 누르시고 다시 접수하여 주시기 바랍니다.
Copyright (c) 2002 by The Korean Society of Medical Informatics