Ontology-Based Question Answering System in Restricted Domain

Abstract

The complexity of natural language presents difficult challenges that traditional Questions and Answers (Q&A) system such as Frequently Asked Questions, relied on the collective predefined questions and answers, unable to address. Traditional Q&A system is unable to retrieve exact answer in response to different kind of natural language questions asked by the user. Therefore, this paper aims to present an architecture of Ontology-based Question Answering (OQA) system, applied to library domain. The main task of OQA system is to parse question expressed in natural language with respect to restricted domain ontology and retrieve the matched answer. Restricted ontology model is designed as a knowledge base to assist the process based on the effective information derived from the questions. In addition, ontology matching algorithm is developed to deal with the questionanswer matching process. A case study is taken from the library of Sultanah Nur Zahirah of Universiti Malaysia Terengganu. A prototype of Sultanah Nur Zahirah Digital Learning ONtologybased FAQ System (SONFAQS) is developed. The experimental result shows that the architecture is feasible and significantly improves man-machine interaction by shortening the searching time.