Ontology-based recommender system of TV programmes for personalisation service in smart TV

Many researches endeavour to solve the problem of information overload and deliver the preferred TV programme content to users. However, they have a weakness in the similarity computation of TV programmes for recommendation because they do not consider the knowledge structures of TV programme content. In this paper, we propose a similarity matching and recommendation method for TV programme content to reduce information overload. Our approach is composed of the three major tasks: (1) conceptualisation and construction of TV programme domain ontologies, (2) computation of the similarity of programme content based on domain ontologies and (3) user preference-based filtering and semantic score-based ranking of the recommendation list. To support semantic-based TV content searching and recommendation, we first design the semantic model of TV ontology and a conceptual process to transform the textual content descriptions of TV programmes to semantic maps. Subjective experiments confirm that the proposed method is effective in semantic-based searching and recommendation.