A semantic similarity measure for recommender systems

In the past few years, recommender systems and semantic web technologies have become main subjects of interest in the research community. In this paper, we present a domain independent semantic similarity measure that can be used in the recommendation process. This semantic similarity is based on the relations between the individuals of an ontology. The assessment can be done offline which allows time to be saved and then, get real-time recommendations. The measure has been experimented on two different domains: movies and research papers. Moreover, the generated recommendations by the semantic similarity have been evaluated by a set of volunteers and the results have been promising.