A Novel Approach to Find Semantic Similarity Measure between Words

Measuring the semantic similarity between words is a significant feature of the web mining domain. The notion of semantic similarity finds applications in various horizons such as relation extraction, community mining, document clustering, and automatic meta data extraction. This paper introduces a method for measuring the semantic similarity of English words. It combines web search engine based similarity measures namely page counts and probability measure from text snippets with the lexical taxonomy based measures of similarity. The adopted measures are employed and learned using support vector machines. The proposed method is successful in achieving a competent accuracy for the said purpose.