The analysis of spatial networks’ evolution has predominantly concentrated on the formation process of links. However, the evolution of networks is similarly shaped by the dissolution of links, which has thus far received considerably less attention. The paper presents separable temporal exponential random graph models (STERGMs) as a promising method in this context, which allows for the disentangling of both processes. Moreover, the applicability of the method to two-mode network data is demonstrated. We illustrate the use of these models for the R&D collaboration network of the German biotechnology industry as well as for testing for the relevance of different forms of proximities for its evolution. The results reveal proximities varying in their relative importance for link formation and link dissolution.

中文翻译：

---

解开空间网络中链路的形成和消解：双模STERGM在德国生物技术行业基于项目的研发网络中的应用

对空间网络演化的分析主要集中在链接的形成过程上。但是，网络的演化同样受到链路分解的影响，到目前为止，受到的关注还很少。本文提出了可分离的时间指数随机图模型（STERGM）作为一种有前景的方法，在这种情况下，这可以使这两个过程脱开。此外，论证了该方法对二模网络数据的适用性。我们举例说明了这些模型在德国生物技术产业的研发合作网络中的使用，以及在测试不同形式的进化对其邻近性的相关性时的使用。结果表明，邻近点在其相对重要的位置对于链环形成和链环溶解而言各不相同。