Exploiting Domain Thesaurus for Medical Record Retrieval
Abstract
InfoLab at the University of Delaware participated in the TREC 2012 Medical Records Track. This paper explains our method and describes experiment results. One limitation of existing keyword matching based retrieval functions is the problem of vocabulary mismatch. To overcome this limitation, we propose to first map topics and visits to bags of concepts using domain thesaurus, and then model the relevance based on the similarities between those concepts.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 2012
- Accession Number
- ADA581309
Entities
People
- Hui Fang
- Miguel A. Callejas P.
- Yue Wang
Organizations
- University of Delaware