An Ontology-Based Question Answering System for University Admissions Advising

Abstract

Question-Answer systems are now very popular and crucial to support human in automatically responding frequent questions in many fields. However, these systems depend on learning methods and training data. Therefore, it is necessary to prepare such a good dataset, but it is not an easy job. An ontology-based domain knowledge base is able to help to reason semantic information and make effective answers given user questions. This study proposes a novel chatbot model involving ontology to generate efficient responses automatically. A case study of admissions advising at the International University–VNU HCMC is taken into account in the proposed chatbot. A domain ontology is designed and built based on the domain knowledge of university admissions using Protégé. The Web user interface of the proposed chatbot system is developed as a prototype using NetBeans. It includes a search engine reasoning the ontology and generating answers to users’ questions. Two experiments are carried out to test how the system reacts to different questions. The first experiment examines questions made from some templates, and the second one examines normal questions taken from frequent questions. Experimental results have shown that the ontology-based chatbot can release meaningful and long answers. The results are analysed to prove the proposed chatbot is usable and promising.