by Ling Song, Jun Ma, Hui Liu, Li Lian, Dongmei Zhang
Abstract:
The main focus of this paper concerns the measuring similarity in a content-based information retrieval and intelligent question-answering environment. While the measure of semantic similarity between concepts based on hierarchy in ontology is well studied, the measure of semantic similarity in an arbitrary ontology is still an open problem. In this paper we define a fuzzy semantic similarity measure based on information theory that exploits both the hierarchical and non-hierarchical structure in ontology. Our work can be generalized the following: firstly each concept is defined as a semantic extended fuzzy set along its semantic paths; secondly the semantic similarity between two concepts is computed with two semantic extended fuzzy sets instead of two concepts themselves. Our fuzzy measure considers some factors synthetically such as ontological semantic relation density, semantic relation depth and different semantic relations, which can affect the value of similarity. Compared with existed measures, this fuzzy similarity measure based on shared information content could reflect latent semantic relation of concepts better than ever.
Reference:
Fuzzy Semantic Similarity Between Ontological Concepts (Ling Song, Jun Ma, Hui Liu, Li Lian, Dongmei Zhang), Chapter in Advances and Innovations in systems computing sciences and software engineering (K Elleithy, ed.), Springer, 2007.
Bibtex Entry:
@incollection{Song2007,
abstract = {The main focus of this paper concerns the measuring similarity in a content-based information retrieval and intelligent question-answering environment. While the measure of semantic similarity between concepts based on hierarchy in ontology is well studied, the measure of semantic similarity in an arbitrary ontology is still an open problem. In this paper we define a fuzzy semantic similarity measure based on information theory that exploits both the hierarchical and non-hierarchical structure in ontology. Our work can be generalized the following: firstly each concept is defined as a semantic extended fuzzy set along its semantic paths; secondly the semantic similarity between two concepts is computed with two semantic extended fuzzy sets instead of two concepts themselves. Our fuzzy measure considers some factors synthetically such as ontological semantic relation density, semantic relation depth and different semantic relations, which can affect the value of similarity. Compared with existed measures, this fuzzy similarity measure based on shared information content could reflect latent semantic relation of concepts better than ever.},
author = {Song, Ling and Ma, Jun and Liu, Hui and Lian, Li and Zhang, Dongmei},
booktitle = {Advances and Innovations in systems computing sciences and software engineering},
editor = {Elleithy, K},
keywords = {SML-LIB-BIBLIO,lang:ENG},
mendeley-tags = {SML-LIB-BIBLIO,lang:ENG},
pages = {275--280},
publisher = {Springer},
title = {{Fuzzy Semantic Similarity Between Ontological Concepts}},
url = {http://www.springerlink.com/index/LT72120883155821.pdf},
year = {2007}
}