A Hybrid Document Recommender Algorithm Based on Random Walk

Abstract

Research shows that recommendations comprise a valuable service for users of a digital library. We proposed a hybrid document recommender system based on random walk. It builds correlation network among users based on the conditional probability in order to solve the sparsity of collaborative filtering. On the other hand, it computes the rating of source user for target item not only based on the neighborhoods’ ratings for target item but also based on the neighborhoods’ ratings for item which is most similar to target item. This can solve the cold start problem of recommender systems. We performed an evaluation on the dataset of National Science and Technology Library. Experimental results illustrate the superiority of the proposed method.