Semantic Similarity between Two Documents: A Topic map Approach

Abstract

Computing semantic similarity between any two entities (word, sentences, documents) is crucial tasks on the web .Semantic Similarity plays a significant and big role in information retrieval(IR), natural language Processing(NLP) and many other tasks of IR related tasks such as relation extraction, and document clustering. It is a concept where a pair of documents is measured to computing the Semantic Similarity between documents using various similarity measures. Computing similarity between a pair of documents with efficient method is really a major difficult task for the user. Similarity measure those are used to find similarity, assign a real number between 0 and 1 to a pair of documents. If both documents are similar then user will get a numerical value 1 otherwise they will get 0.This paper proposes a framework for computing the semantic similarity between documents based on topic maps. The process starts with pre-processing of the documents using NLP parser. Then Topic map is build that represent the document in compact form and cosine similarity measures is used to measure the similarity between these topic maps.