SCIENCE@home

Discussion of “Co-citation and Co-authorship Networks of Statisticians”

Joshua Daniel Loyal,Yuguo Chen

Published 2022 in Journal of business & economic statistics

ABSTRACT

We want to congratulate the authors on a fascinating article containing an insightful analysis and their hard work curating the high-quality co-citation and co-authorship networks. These datasets alone are a valuable contribution to the statistics profession, which will undoubtedly inspire future data science projects and advances in methodology. In fact, we are eager to use these networks in our own classrooms and research. Furthermore, the authors use these networks to tackling exciting questions in network science that go beyond the familiar problems of edge imputation and predicting node labels. In doing so, the authors perform a terrific analysis accompanied by exciting new methodology. This analysis serves as a great first step in understanding these networks, and the ideas initiated in this article will certainly stimulate many further research questions. For example, how do individuals influence the research trajectory of others? Or, how do the components of the proposed “research map” change over time? As statisticians, we have a first-hand understanding of the complex system these networks describe, which can help us contextualize these problems and validate our inferences. As such, we look forward to this dataset becoming a standard benchmark to test new models and scalable inference procedures. A central challenge of the work is rigorously quantifying the time-varying research patterns and trends of the statistics community, which naturally leads to the statistical modeling of dynamic networks. The authors skillfully use various dynamic block models to uncover statisticians’ community structure. In the remainder of this discussion, we focus on an alternative statistical network model known as latent space models. Specifically, we briefly describe the latent space modeling approach, highlight five further research questions, and demonstrate how latent space models may be used to answer them. Although other models, such as block models, may be appropriate to tackle these questions as well, we hope that this discussion gives future researchers an expanded toolset to investigate this rich data source. Latent space models (LSMs) are a popular approach to modeling networks first proposed by Hoff, Raftery, and Handcock (2002) for static networks and later generalized to dynamic networks by Sarkar and Moore (2006) and Sewell and Chen (2015). These models embed the nodes of a network into a low-dimensional latent space, which can provide meaningful

PUBLICATION RECORD

CITATION MAP

EXTRACTION MAP

CLAIMS

  • No claims are published for this paper.

CONCEPTS

  • No concepts are published for this paper.

REFERENCES

Showing 1-10 of 10 references · Page 1 of 1

CITED BY