Header

UZH-Logo

Maintenance Infos

Absence of a resolution limit in in-block nestedness


Mariani, Manuel; Pallazi, Maria J; Solé-Ribalta, Albert; Borge-Holthoefer, Javier; Tessone, Claudio (2021). Absence of a resolution limit in in-block nestedness. Communications in Nonlinear Science and Numerical Simulation, 94:105545.

Abstract

Nestedness refers to a hierarchical organization of complex networks where a given node’s neighbors tend to form a subset of the neighborhoods of higher-degree nodes. Although nestedness has been traditionally interpreted as a macroscopic property that involves all the nodes of the network, recent works have reinterpreted it as a mesoscopic network property, by revealing that interactions in diverse empirical networks are often arranged into blocks that exhibit an internally nested structure. Inspired by the popular modularity function, these works rely on a quality function – called in-block nestedness – that assumes a partition of the nodes into blocks that exhibit an internal nested structure. A potential limitation of this approach is that the optimization of modularity (and related quality functions) notoriously suffers from resolution limits that impair the detectability of small blocks. Yet, we do not know whether the in-block nestedness function may exhibit similar resolution limits. Here, we provide numerical and analytical evidence that the in-block nestedness function lacks a resolution limit, which implies that our capacity to detect correct partitions in networks via its maximization depends solely on the accuracy of the optimization algorithms.

Abstract

Nestedness refers to a hierarchical organization of complex networks where a given node’s neighbors tend to form a subset of the neighborhoods of higher-degree nodes. Although nestedness has been traditionally interpreted as a macroscopic property that involves all the nodes of the network, recent works have reinterpreted it as a mesoscopic network property, by revealing that interactions in diverse empirical networks are often arranged into blocks that exhibit an internally nested structure. Inspired by the popular modularity function, these works rely on a quality function – called in-block nestedness – that assumes a partition of the nodes into blocks that exhibit an internal nested structure. A potential limitation of this approach is that the optimization of modularity (and related quality functions) notoriously suffers from resolution limits that impair the detectability of small blocks. Yet, we do not know whether the in-block nestedness function may exhibit similar resolution limits. Here, we provide numerical and analytical evidence that the in-block nestedness function lacks a resolution limit, which implies that our capacity to detect correct partitions in networks via its maximization depends solely on the accuracy of the optimization algorithms.

Statistics

Citations

Altmetrics

Downloads

6 downloads since deposited on 20 Oct 2020
0 downloads since 12 months
Detailed statistics

Additional indexing

Item Type:Journal Article, refereed, original work
Communities & Collections:03 Faculty of Economics > Department of Business Administration
08 Research Priority Programs > Social Networks
Dewey Decimal Classification:330 Economics
Language:English
Date:March 2021
Deposited On:20 Oct 2020 16:20
Last Modified:24 Nov 2023 02:42
Publisher:Elsevier
ISSN:1007-5704
OA Status:Closed
Publisher DOI:https://doi.org/10.1016/j.cnsns.2020.105545
Other Identification Number:merlin-id:19830