Interest in the network analysis of bibliographic data has increased significantly in recent years. Yet, appropriate statistical models for examining the full dynamics of scientific citation networks, connecting authors to the papers they write and papers to other papers they cite, are not available. Very few studies exist that have examined how the social network between co-authors and the citation network among the papers shape one another and co-evolve. In consequence, our understanding of scientific citation networks remains incomplete. In this paper we extend recently derived relational hyperevent models (RHEM) to the analysis of scientific networks, providing a general framework to model the multiple dependencies involved in the relation linking multiple authors to the papers they write, and papers to the multiple references they cite. We demonstrate the empirical value of our model in an analysis of publicly available data on a scientific network comprising millions of authors and papers and assess the relative strength of various effects explaining scientific production. We outline the implications of the model for the evaluation of scientific research.

翻译：近年来，文献数据的网络分析引起了显著增长的兴趣。然而，目前尚缺乏适当的统计模型来全面研究科学引文网络的动态机制，包括连接作者与其所著论文、以及论文与其所引文献之间的关系。极少有研究探讨合著者之间的社会网络与论文之间的引文网络如何相互塑造并共同演化。因此，我们对科学引文网络的理解仍不完整。本文将近期提出的关系超事件模型（RHEM）扩展至科学网络分析，提供了一个通用框架来建模涉及多作者与所著论文、以及论文与所引多篇参考文献之间的多重依赖关系。我们通过分析包含数百万作者和论文的公开科学网络数据，实证展示了该模型的价值，并评估了解释科学生产的各种效应的相对强度。最后，我们概述了该模型对科学研究评估的意义。