An ontology-based semantic similarity metric to empower semantic search

Heterogeneity in documents is a challenge for information Retrieval. The keyword search focuses on matching the keywords with web repositories. It does not consider the synonyms or semantically similar words. The heterogeneity of the content makes retrieval inadequate. Semantic search helps to capture more appropriate results using domain ontology. Keyword search is extended with the help of similar concepts of ontology. Similarity between the ontological concepts is recognized to get appropriate search results. Once the semantic similarity among the concepts is known, more relevant documents can be retrieved. In this paper, we propose a metric based on traditional methods, combined with computational techniques to measure the similarity between concepts. The paper gives the concept of DOT (Domain Ontology Tree). It uses conventional definitions of the Tree (Data Structure) for ontology and proposes a method of partitioning to calculate the similarity. The method is based on IS-A hierarchical relationship. We have implemented a prototype system for the support of the proposed method, and also compared it with existing methods, the results are encouraging.