ISPREC: Integrated Scientific Paper Recommendation using heterogeneous information network

Due to the rapid expansion of online scientific articles, researchers have got into trouble finding reliable articles that are relevant to their research interests. Recently, a group of scientific paper recommendation algorithms has been proposed to solve this issue. But, they have two main shortcomings. First, they can only recommend papers to experienced researchers who have published some papers and not amateur ones. Second, they ignore some valuable sources of information in scientific article libraries. This paper presents a novel Integrated Scientific Paper RECommendation approach, called ISPREC, which integrates different pieces of information as a novel heterogeneous network structure, called SPIN. Thereafter, exploits a limited random-walk algorithm for a Top-N recommendation. Extensive experiments on a real-world dataset demonstrate a significant improvement of the proposed framework of ISPREC compared to the state-of-the-art scientific paper recommendation algorithms.