Architecture and design of a novel tool for intelligent general purpose question answering system

摘要

Question answering systems (QAS) requires bonding/direct connectivity between Artificial Intelligence (AI), Knowledge Representation (KR), Information Retrieval (IR), Information Extraction (IE), Reasoning, etc. Now-a-days, QAS is applied to every walk of life like designing expert systems, semantic web, robotics, auto answering machine, etc. Research in the area of QAS provides an interface to the user for IR and IE. The objective of this QAS system is to provide an interactive environment with the real user. User can select the question from the pool of questionnaire and put up to.QAS on real-time basis also. QAS is an intelligent system that allows the user to ask simple as well as complex questions and gives the result not as link or paragraph but returns the exact answer without delay. System aimed to design the general QAS by using the Semantic Network and Script technique to acquire the knowledge and Natural Language Processing (NLP) to process the incoming knowledge. The idea was to improve the system performance and the retrieval efficiency. The implemented system was tested using 100 stored questions and 250 newly entered questions applied on different input collected from various sources like newspaper, internet and book, etc. The result of the QAS shows that the use of Semantic Network, Script and NLP could greatly improve the system performance.