Long Tail Recommender Utilizing Information Diffusion Theory

Our approach aims to provide a mechanism for recommending long tail items to knowledge workers. The approach employs collaborative filtering using browsing features of identified key population of the diffusion of information. We conducted analytic experiment for a novel recommendation algorithm based on the browsing features of identified selected users and discovered that the first 10 users accessing a particular page play the key role in information spread. The evaluation indicated that our approach is effective for long tail recommendation.