PaperRank: A Ranking Model for Scientific Publications

The idea of HITS and PageRank models, which are well known in Web mining, can be used in ranking scientific publications by noting the similarity of World Wide Web and the scientific citation networks. PaperRank, an extension of HITS and PageRank model, is introduced in this paper to evaluate the scientific publications. It retrieves the indirect relationships between papers using relativity measurements in stead of the simple direct citations between papers. The new method provides more valuable knowledge by controlling the relativity functions and weights. The experimental results on a large set of papers in biology show that the proposed method can find more authoritative papers than those approaches based on the traditional methods. The PaperRank can be considered as an efficient complementary tool for ranking scientific publications.