InfoSlim: An Ontology-Content Based Personalized Mobile News Recommendation System

This paper proposes a novel personalized news recommendation system named InfoSlim. The new system uses semantic technique to annotate news items and user preference in order to add rich metadata information into traditional keyword vector. By doing this, the similarity measure between item profile and user profile can be done by not only lexical-level cosine-based method but also by semantic-level ontology-based method. Such recommendation method can efficiently improve the accuracy of recommendation and therefore can better reflect user´s interest and save mobile resources.