An Automated Question Answering(Q/A)System For Academic Environment

Abstract

Almost every Internet user has daily many generals enquires. Users can solve these queries from the Internet Search Engines (ISE). These search engines provide the articles or web pages as result. Users need to expend time on the Internet to get the exact result. Sometimes users may also physically visit the location to get resolve their queries. The academic domain has applicants and students who have more queries about academics. These both have lots of queries and ISEs can't solve these queries. Ontology is used to represent domain-specific data. In this study, Ontology has the format of functional-style syntax. Using Ontology with Semantic Web Rule Language (SWRL) technology. We introduce a domain-specific general Academic Question Answering (Q/A) System for academics. This Q/A System can cover most of the important departments of academics. A user can ask from the system through the submission of data values for a query. Academia Ontology is generated through the Protégé tool. The answer data is retrieved from data properties for applying the relevant rules. Rules are generated with Semantic Web Rule Language (SWRL). This is open to getting focused on Academics Q/A Systems and further applying to other domains with the Semantic Web and Ontology-based approaches. The experimental results explain the feasibility of the Q/A System. This system can respond with more accurate and exact results with an F-Measure of 48%. This Academic Q/A System has the potential to solve the general queries of most of the stakeholders in academia. This Q/A System has been produced to answer a variety of questions in academia. This system can further be extended to embed a natural language interface by employing Natural Language Processing. Further work includes the extension of systems to other domains as well e.g. financial institutions, and government bodies.