Using ontology for measuring semantic similarity for question answering system

Abstract

Semantic similarity is an essential concept that widen across various fields such as artificial intelligence, natural language processing, information retrieval, relation extraction, document clustering and automatic data extraction. The study proposed in this paper focus on semantic similarity and counter measure on question answering task. Despite the usefulness of semantic similarity in these applications, finding the exact meaning between the words from question and answer pair has become a major challenge in question answering system. To recognize various relationships those exist between question and answer pair, measuring semantically is essential. We present study using a match maker algorithm for similarity measurement between question-answer pair. The study compares similarity measurement using match maker algorithm as against semantic similarity measure on Miller-Charles bench marked data set. This paper present study about different types of algorithm used to measure the semantic similarity between the words and those result compare with match maker algorithm. The experiments on real datasets shows matchmaker algorithm works better than web-based semantic similarity measure.