关键词: 文献, 相似性, 科学引用, 指数随机图模型

Abstract: [Purpose/significance] It is possible to understand the mechanism of citation more deeply and clearly, understanding the nature of phenomenon in similarity between citing document and cited document which share the same or similar features.[Method/process] Aiming to discover the extent to which similarity between documents existing citation relationship affects the preference of citations, empirical research was conducted to focus on the academic field of LIS by the statistical method of Exponential Random Graph Models (ERGM). [Result/conclusion] Some empirical results were found as following that there obviously exists tendency to be similarity between documents in the aspects of embedded network structure, affiliation and journal. Specially, the dyadic citation relation would be more likely to be embedded in the triangle transmit structure in citation network, and happened between the documents with the same affiliation and journal. Moreover, the documents, from the countries which are domain in the specific academic field, would be more likely to form the relationship of citation. Consequently, the empirical results adequately imply that social affinity is the crucial mechanism for citation behavior and reflect the social feature of citation preference.

Key words: documents, similarity, scientific citation, ERGM