Question classification for medical domain Question Answering system

Abstract

Question classification plays an important role in question answering system. It helps in finding or constructing accurate answers and hence improves the quality of Question Answering systems. The question classification approaches generally used are: Rule based, Machine learning and Hybrid. This paper presents our research work on question classification through rule based approach. The question processing module helps in assigning a suitable question category and identifying the keywords from the given input question. A prototype system based on the proposed method has been constructed and the experiment on 500 medical questions collected from patients and doctors has been carried out. Using the two layered taxonomy of 6 course grain and 50 fine grained categories developed by Li and Roth, we have classified the questions into various categories. We have also studied the syntactic structure of the question and suggest the syntactic patterns for particular category of questions. Using these question patterns we have classified the question into particular category. In this paper we have proposed a compact and effective method for question classification. The experimental output shows that even with small set of question categories we can classify the questions with more satisfactory and better result.