Word/phrase based answer type classification for Bengali question answering system

Abstract

This paper work demonstrated an analytical ability of question answer type classification steps towards building a question answering system in Bengali language. In order to respond correctly to a free form factoid question or complex questions given a large collection of corpus, one needs to understand the question category that allows determining some of the constraints the question imposes on a possible answer. Question classification (QC) Systems allows the user to ask questions in a natural language and acquire a correct answer category (AC). This work presents the first work on a machine learning approach to Bengali question classification using stochastic gradient descent (SGD) classifier method. Stochastic gradient descent is a gradient descent optimization method for minimizing an objective function that is written as a sum of differentiable functions. Our system based on stochastic gradient descent (SGD) classifier achieved average precision 0.95562 for coarse and 0.87646 for finer.