An ontology-based method for personalized recommendation

Abstract

Recommender systems aim to provide users with personalized suggestions about information items, products or services that are likely to be of their interests. The traditional syntactic-based recommender systems suffer from a number of shortcomings for the information available on the internet has been designed to be readable only by humans and computer systems can not effectively process nor interpret the data present in it. As one Semantic Web technology, Ontology facilitates the knowledge sharing, reuse, communication, collaboration and construction of knowledge rich and intensive systems. Adding semantically empowered techniques to recommender systems can significantly improve the overall quality of recommendations. In this paper, an ontology-based method for personalized recommendation of knowledge in the heterogeneous environment is presented, which provides users with an autonomous tool that is able to minimize repetitive and tedious retrieved information. It constructs a domain ontology by integrating multi-resource and heterogeneous data, generates a user's interest ontology by analyzing the user's demographic characteristics and personal preferences. Based on the matching results of the domain ontology, user's query requests and interest ontology, the recommender system can suggest the proper information to the user who is likely interested in the related topics. Finally, qualitative evaluation is carried out in order to demonstrate the effectiveness of the proposed method.