Ontology-Based Interactive Question and Answering System

Abstract

Interactive Question and Answering System (IQA) is a new question and answer research field. IQA guides users to complement the missing question information through some interaction strategy. It overcomes the defect that traditional question answering system cannot handle questions with incomplete information; it also reduces the requirement for questions analysis. This paper introduces the mechanism of IQA, proposes a domain Ontology-based IQA framework and an incomplete questions' knowledge representing framework-question information domain (QID). Domain Ontology-based IQA extracts effective information from domain question corpus, and then focuses the effective information on Ontology concept through Ontology-based query expansion (OBQE) method, to establish Ontology-based domain QID knowledgebase. The interaction is, searching the domain QID knowledgebase to complement users' QID to domain QID. We use rough set algorithm to deal with domain QID knowledgebase, reduce the search space and interaction steps. The experiment result shows that our IQA framework is feasible and the reduction of knowledgebase improves the interaction efficiently.