Concept similarity computation based on ontology

Finding a good similarity algorithm for the use on ontology is important to some techniques such as semantic retrieval, data mining, ontology translations etc. Similarity computation has special characters when applied in different domains. However after we learned most of the similarity measures based on ontology, we found that most researchers did not deal with these characters. In this paper we work at the similarity computation based on ontology applied in semantic retrieval. We analyze the characters about this application and problems to be improved, and then we propose an improved semantic similarity computation model. At last, we demonstrate the improvement of this model through experiments.