Using a semantic multidimensional approach to create a contextual recommender system
Item recommendations calculated by recommender systems mostly in use today, only rely on item content description, user feedback and profile information. In modern mobile services, however, contextual information and semantic knowledge can play a significant role concerning the quality of these recommendations. Therefore, the SMART Recommendations Engine of Fraunhofer FOKUS is extended by the SMART Multidimensionality Extension and the SMART Ontology Extension that enable the recommender to incorporate contextual and semantic data into the recommendation process. The demonstration of the SMART Ontology Extension visualizes that the preciseness of recommendations can be increased by exploiting implicit and indirect knowledge, classification and location information gained from ontologies when generating recommendations in the scope of an exemplary food purchase scenario.