Estimating the Number of Communities in a Network

Newman, M. E. J.; Reinert, Gesine

doi:10.1103/physrevlett.117.078301

Cited by 127 publications

(127 citation statements)

References 33 publications

(65 reference statements)

Supporting

Mentioning

127

Contrasting

Order By: Relevance

“…A major drawback, however, of overlapping community detection algorithms is that the number of communities within a network needs to be known in advance (Ball et al, 2011). Typically, this number is unknown, although recent studies have attempted to apply Bayesian inference and Monte Carlo methods to estimate this number (Newman & Reinert, 2016;Riolo, Cantwell, Reinert, & Newman, 2017). However, a successful application of such methods highly depends on the choice of an appropriate prior probability.…”

Section: Community Detection Algorithmmentioning

confidence: 99%

Mapping the global network of fisheries science collaboration

et al. 2019

View full text Add to dashboard Cite

As socio‐environmental problems have proliferated over the past decades, one narrative which has captured the attention of policymakers and scientists has been the need for collaborative research that spans traditional boundaries. Collaboration, it is argued, is imperative for solving these problems. Understanding how collaboration is occurring in practice is important, however, and may help explain the idea space across a field. In an effort to make sense of the shape of fisheries science, here we construct a co‐authorship network of the field, from a data set comprising 73,240 scientific articles, drawn from 50 journals and published between 2000 and 2017. Using a combination of social network analysis and machine learning, the work first maps the global structure of scientific collaboration amongst fisheries scientists at the author, country and institutional levels. Second, it uncovers the hidden subgroups—here country clusters and communities of authors—within the network, detailing also the topical focus, publication outlets and relative impact of the largest fisheries science communities. We find that whilst the fisheries science network is becoming more geographically extensive, it is simultaneously becoming more intensive. The uncovered network exhibits characteristics suggestive of a thin style of collaboration, and groupings that are more regional than they are global. Although likely shaped by an array of overlapping micro‐ and macro‐level factors, the analysis reveals a number of political–economic patterns that merit reflection by both fisheries scientists and policymakers.

show abstract

Section: Community Detection Algorithmmentioning

confidence: 99%

Mapping the global network of fisheries science collaboration

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Other techniques to extract the number of groups have been proposed (Côme and Latouche, 2015;Daudin et al, 2008;Handcock et al, 2007;Latouche et al, 2012;Newman and Reinert, 2016).…”

Section: E Methods Based On Statistical Inferencementioning

confidence: 99%

Community detection in networks: A user guide

2016

View full text Add to dashboard Cite

Community detection in networks is one of the most popular topics of modern network science. Communities, or clusters, are usually groups of vertices having higher probability of being connected to each other than to members of other groups, though other patterns are possible. Identifying communities is an ill-defined problem. There are no universal protocols on the fundamental ingredients, like the definition of community itself, nor on other crucial issues, like the validation of algorithms and the comparison of their performances. This has generated a number of confusions and misconceptions, which undermine the progress in the field. We offer a guided tour through the main aspects of the problem. We also point out strengths and weaknesses of popular methods, and give directions to their use.

show abstract

“…The stochastic block model assumes that each node belongs to one of p latent communities, and the probability of connection between two nodes is given by a p × p connectivity matrix. This model has been extended in various directions, by introducing degree correction parameters (Karrer and Newman, ), by allowing the number of communities to be unknown or to grow with the size of the network (Kemp et al ., ; Newman and Reinert, ) or by considering overlapping communities (Airoldi et al ., ; Miller et al ., ; Latouche et al ., ; Palla et al ., ; Yang and Leskovec, ). Stochastic block models and their extensions have shown that they offer a very flexible modelling framework, with interpretable parameters, and have been successfully used for the analysis of numerous real world networks.…”

Section: Introductionmentioning

confidence: 99%

Exchangeable Random Measures for Sparse and Modular Graphs with Overlapping Communities

Todeschini

Caron²

2020

Journal of the Royal Statistical Society Series B: Statistical Methodology

View full text Add to dashboard Cite

Summary We propose a novel statistical model for sparse networks with overlapping community structure. The model is based on representing the graph as an exchangeable point process and naturally generalizes existing probabilistic models with overlapping block structure to the sparse regime. Our construction builds on vectors of completely random measures and has interpretable parameters, each node being assigned a vector representing its levels of affiliation to some latent communities. We develop methods for efficient simulation of this class of random graphs and for scalable posterior inference. We show that the approach proposed can recover interpretable structure of real world networks and can handle graphs with thousands of nodes and tens of thousands of edges.

show abstract

Estimating the Number of Communities in a Network

Cited by 127 publications

References 33 publications

Mapping the global network of fisheries science collaboration

Mapping the global network of fisheries science collaboration

Community detection in networks: A user guide

Exchangeable Random Measures for Sparse and Modular Graphs with Overlapping Communities

Contact Info

Product

Resources

About