Question Answering System in Bengali Using Semantic Search

A Semantic Question Answering system (QA) is a computer program that takes question as input and retrieves answer using Information Retrieval (IR) rather than keyword based searching mechanism. Currently, there exists several Question Answering approaches and most of them are lexical similarity based approach. This work presents the design, implementation and evaluation of QA system based on semantic analysis using Bengali WordNet. The data sets are prepared from the Bengali text corpus which is developed under the TDIL (Technology Development for Indian Languages) project of the Government of India. The proposed model is tested on 40 questions from different subject domains and the overall accuracy is achieved as 80%. The challenges and the pitfalls of this work are explained in detail at the end of this paper.