Tourism Destination Recommendation Using Ontology-based Conversational Recommender System

Abstract

Traveling is one of the human psychological needs. Many choices and lack of information about desired tourist attractions are some of the obstacles in fulfilling the need. One of the technologies to overcome these obstacles is the recommender system which can provide recommendations for the users to choose some interesting tourist attractions from several tourist destinations. A conversational recommender system (CRS) offers a way of recommending tourist destinations in a conversational mechanism. We use ontology as a representation of knowledge to generate conversational interactions, recommendations, and explanation facilities. With this ontology-based CRS, we can overcome cold start problems, and also the system can guide the users to get the desired tourist attractions. In this study, we use a combination of navigation by asking (NBA) and navigation by proposing (NBP) strategy to generate interactions on the CRS. Based on the user study, in general, users find it helpful in finding tourist destinations that suit their needs. It is because the factor of trust and perceived ease of use comes from the interaction and explanation facilities contained in the system. Besides, perceived usefulness directly affects users' interest in utilizing this CRS interaction model in the future.