Web Mining for Semantic Similarity between Words

Semantic similarity is a context dependent and dynamic phenomenon. A train is semantically similar to a horse if the context is moving objects. On the otherhand if the context is living beings the same objects may not have any similarity. Finding semantic similarity has increasingly become important in many applications such as community mining, data clustering, relation extraction, query expansion and many natural language processing projects. The paper addresses the general framework for finding semantic similarity between words and its utility in the state of art web applications. It presents some empirical results using page count and snippet based methods on the web.