Home | DTI | 2004–05 funded proposals | Ted Pedersen, Serguei Pahkomov
Ted Pedersen, Serguei Pahkomov
Developing Measures of Semantic Relatedness for the Biomedical Domain
The proposed research will result in measures that will automatically quantify the extent to which two biomedical concepts are similar or related to each other. We will evaluate the suitability of generalized measures of semantic similarity and relatedness when applied to concepts found in two widely used ontologies in the biomedical domain, MeSH and SNOMED. We hypothesize that existing measures will be unable to make the fine grained distinctions found in these resources, since they have been developed for concepts as found in general English text. We will adapt and extend one or more of these measures to the biomedical domain through the use of statistical information and lexical chains derived from large corpora of biomedical texts. The proposed research will be carried out in collaboration with Dr. Serguei Pahkomov of the Division of Medical Informatics at the Mayo Clinic in Rochester, Minnesota.
Digital Technology Center
499 Walter Library, 117 Pleasant Street SE, Minneapolis, MN 55455
P: (612) 624-9510