The goal of this fellowship will be to design and test an algorithm and write a program to automate mapping SNOMED International veterinary terminology to MeSH synonyms to facilitate subject searching for veterinary literature contained in the National Library of Medicine Medline database. Fur the purpose of this application simple mapping refers to finding a path from a work in one medical vocabulary, in this case SNOMED International, to a synonym or synonyms in another vocabulary, such as MeSH. The small animal veterinary disease, diagnosis terms from SNOMED International will be used as the testbed. The following hypotheses will be tested: 1) It is possible to develop an algorithm to map fine-grained SNOMED veterinary terminology to synonyms in the form of qualified MeSH headings that will retain the meaning of the original veterinary term. 2) Simple mapping of the SNOMED veterinary terminology to qualified MeSH headings will improve precision of information retrieval over that of simple keyword searching. Evaluations will be performed on the algorithms ability to correctly map the SNOMED testbed, and the ability to retrieve precise documents using the mapped MeSH terms with qualifiers.
