Improving content based recommender systems using linked data cloud and FOAF vocabulary

With the deluge of data published on the web, it becomes even more difficult for a user to get access to the relevant information based on his preferences. In order to accurately predict the preference a user would give to an item, recommender systems should use an effective information filtering engine. This task can be achieved using content based filtering (CBF) or collaborative filtering or a hybrid approach. This work describes an approach to CBF that aims to deal with the issues of unstructured data and new user on which existing approaches perform poorly. The basic feature of the proposed approach is to incorporate linked data cloud into the information filtering process using a semantic space vector model. FOAF vocabulary is used to define a new distance measure between users based on their FOAF profiles. Unstructured items representations are enhanced by additional attributes extracted from Linked data cloud which alleviates the burden to analyze the content of these items and therefore reduces the computational cost. We report on some promising experiments of the proposed approach performed on MovieLens data sets.