Ontology based Semantic Similarity Measure using Concept Weighting
S. Anitha Elavarasi, J. Akilandeswari, K. Menaga · 2014
Semantic similarity between the documents is essential when it is extracted from free text document. Representing the presence and absence of concept in binary format may not provide perfect accuracy. Concept weighting through term frequency will increase accuracy of clustered document. Concept weight is determined using term frequency and semantic distance. Semantic similarity of a concept is derived using ontology extracted from swoogle. Vector space model with parent-child (is-a) relationship ontology are exploited using protege. Term frequencies for the extracted concepts are calculated using text processing. In this paper Cosine similarity using concept weight measure is applied to find similarity between different documents. According to the similarity score, documents are clustered. In this paper a sample walkthrough for the proposed system has been discussed by comparing two documents.