A Novel Personalized Academic Knowledge Sharing System in Online Social Network

Abstract

Information overload is a major problem for both readers and authors due to the rapid increase in scientific papers in recent years. Methods are proposed to help readers find right papers, but few research focuses on knowledge sharing and dissemination from authors’ perspectives. This paper proposes a personalized academic knowledge sharing system that takes advantages of author’s initiatives. In our method, we combine the user-level and document-level analysis in the same model, it works in two stages: 1) user-level analysis, which is used to profile users in three dimensions (i.e., research topic relevance, social relation and research quality); and 2) document-level analysis, which calculates the similarity between the target article and reader’s publications. The proposed method has been implemented in the ScholarMate, which is a popular academic social network. The experiment results show that the proposed method can effectively promote the academic knowledge sharing, it outperforms other baseline methods.