Applying text similarity algorithm to analyze the triangular citation behavior of scientists

Abstract

Triangular citation is a particular citation structure with research value and application significance, and the fundamental mechanism of its formation lies in the indirect citation between Literature A and Literature C. It is necessary to dig deeply into the citation context and citation motivation of indirect citation behavior to understand the internal mechanism of triangular citation. In this paper, the text-similarity algorithm was used to calculate the similarity of citation contents between B→A and C→A in 22,536 triangular citation data, and the followers C with high citation similarity were identified. Secondly, from the perspectives of language, literature type, interdisciplinary citation, and author self-citation, combined literature characteristics and followers’ identification results, the context and motivation of indirect citation behaviors imposed by the followers C were analyzed. The results show that followers C appear more frequently in triangular citations. At the same time, due to the influence of language differences, literature types, discipline differences or author self-citation, follower C will have lazy citation motivation when citing Literature A, partially or out-and-out requote Literature A from the citation content of Literature B, resulting in a high similarity between the citation content of C→A and B→A.