by Angelos Hliaoutakis
Abstract:
Semantic Similarity relates to computing the similarity between concepts, having the same meaning or related information, which are not lexicographically similar. This is an important problem in Natural Language Processing and Information Retrieval Research and has received considerable attention in the literature. Several algorithmic approaches for computing semantic similarity have been proposed. We investigate approaches for computing semantic similarity by mapping terms or con- cepts to an ontology and by examining their relationships in that ontology. Comparing concepts that belong to different ontologies is far more difficult problem. Some of the most popular semantic similarity approaches are implemented and evaluated based on WordNet as the underlying reference ontology. We also propose a method for comparing terms in different ontologies. In this work we examined similarity between terms basically from MeSH (medical) and WordNet ontologies. Building upon the idea of semantic similarity we also propose an information re- trieval methodology capable of detecting similarities between documents containing semantically similar but not necessarily identical terms. Our proposed Information Retrieval model has been evaluated for retrieval of documents in Medline database. The experimental results demonstrated that our proposed model (although slower) achieves significant performance improvements compared to the state-of-the-art approach based on the Vector Space Model.
Reference:
Semantic similarity measures in MeSH ontology and their application to information retrieval on Medline (Angelos Hliaoutakis), PhD thesis, Technical University of Crete, Greek, 2005.
Bibtex Entry:
@phdthesis{Hliaoutakis2005,
abstract = {Semantic Similarity relates to computing the similarity between concepts, having the same meaning or related information, which are not lexicographically similar. This is an important problem in Natural Language Processing and Information Retrieval Research and has received considerable attention in the literature. Several algorithmic approaches for computing semantic similarity have been proposed. We investigate approaches for computing semantic similarity by mapping terms or con- cepts to an ontology and by examining their relationships in that ontology. Comparing concepts that belong to different ontologies is far more difficult problem. Some of the most popular semantic similarity approaches are implemented and evaluated based on WordNet as the underlying reference ontology. We also propose a method for comparing terms in different ontologies. In this work we examined similarity between terms basically from MeSH (medical) and WordNet ontologies. Building upon the idea of semantic similarity we also propose an information re- trieval methodology capable of detecting similarities between documents containing semantically similar but not necessarily identical terms. Our proposed Information Retrieval model has been evaluated for retrieval of documents in Medline database. The experimental results demonstrated that our proposed model (although slower) achieves significant performance improvements compared to the state-of-the-art approach based on the Vector Space Model.},
author = {Hliaoutakis, Angelos},
booktitle = {2005-11-01).[2007-12-10]. http://www. \ldots},
keywords = {SML-LIB-BIBLIO,lang:ENG},
mendeley-tags = {SML-LIB-BIBLIO,lang:ENG},
school = {Technical University of Crete, Greek},
title = {{Semantic similarity measures in MeSH ontology and their application to information retrieval on Medline}},
type = {Master’s thesis},
url = {http://www.intelligence.tuc.gr/publications/Hliautakis.pdf},
year = {2005}
}