Question Classification using syntactic and rule based approach

Question Classification is a vital component of Question Answering System. In this paper we have proposed a compact and effective method for question classification. Here rather than using a two layered taxonomy of 6 course grain and 50 fine grained categories developed by Li and Roth, 2002, we have classified the questions into three broad categories. We have also studied the syntactic structure of the question and suggest the syntactic patterns and expected answer type for particular category of questions. Using these question Patterns we have also suggested an algorithm for classifying the question into particular category. For experiment purpose we have used Li and Roth data set of 2000 questions. The experimental output shows that even with small set o question categories we can classify the questions with more satisfactory and better result. Hence in brief various contributions through this paper are: - Achieve state of art of question classification using less number of question categories - Communicate an Question classification algorithm which classifies the question into proposed categories which aids to embed the appropriate Answer Extraction Algorithm discussed in our previous work [23]. - Suggest more generic syntactic patterns for Wh questions.