论文标题
一种多层网络方法，用于建模作者身份对物理期刊引用动态的影响
A multi-layer network approach to modelling authorship influence on citation dynamics in physics journals
论文作者
论文摘要
我们提供一个通用框架，以建模由不同耦合层组成的网络的增长。我们的目的是估计一个这样的层对其他层的动态的影响。作为应用程序，我们研究了一个科学计量学网络，其中一层由出版物作为节点和引用作为链接组成，而第二层则代表作者。这允许解决以下问题，例如作者的特征，例如其出版物数量或以前的合着者数量，会影响新出版物的引文动态。为了测试有关这种影响的不同假设，我们的模型以不同的方式结合了引文成分和社会成分。然后，我们评估它们在复制九种不同物理期刊中引文动态方面的性能。为此，我们为统计参数估计和模型选择开发了一种通用方法，该方法适用于增长的多层网络。它同时考虑了参数误差和模型复杂性，并且在计算上有效且可扩展到大型网络。
We provide a general framework to model the growth of networks consisting of different coupled layers. Our aim is to estimate the impact of one such layer on the dynamics of the others. As an application, we study a scientometric network, where one layer consists of publications as nodes and citations as links, whereas the second layer represents the authors. This allows to address the question how characteristics of authors, such as their number of publications or number of previous co-authors, impacts the citation dynamics of a new publication. To test different hypotheses about this impact, our model combines citation constituents and social constituents in different ways. We then evaluate their performance in reproducing the citation dynamics in nine different physics journals. For this, we develop a general method for statistical parameter estimation and model selection that is applicable to growing multi-layer networks. It takes both the parameter errors and the model complexity into account and is computationally efficient and scalable to large networks.
