Mapping change in large networks - bootstrap resampling and significance clustering

Oct 26, 2026

Upload by Adam

Change is a core feature of biological, technological, economic, and scientific interactions, yet existing network-clustering tools often cannot separate real structural shifts from noise in single observed networks. The paper presents a framework combining bootstrap resampling with significance clustering to assign statistical meaning to partitions. It then summarizes significant structural changes using alluvial diagrams and applies the method to citation networks of about 7000 journals, showing how neuroscience evolved over the past decade.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Introduction

Need to quantify change in large networks

Results

Journal Citation Networks

arXiv : 0812 . 1242v2 [physics .soc-ph] 6 Oct 2010 Mapping change in large networks M. Rosvall􀀃 Department of Biology, University of Washington, Seattle, Washington, United States of America C. T. Bergstrom Department of Biology, University of Washington, Seattle, Washington, United States of America and Santa Fe Institute, Santa Fe, New Mexico, United States of America (Dated: October 22, 2018) Change is a fundamental ingredient of interaction patterns in biology, technology, the economy, and science itself: Interactions within and between organisms change; transportation patterns by air, land, and sea all change; the global 􀀌nancial 􀀍ow changes; and the frontiers of scienti􀀌c research change. Networks and clustering methods have become important tools to comprehend instances of these large-scale structures, but without methods to distinguish between real trendsand noisy data, these approaches are not useful for studying how networks change. Only if we can assign signi􀀌cance to the partitioning of single networks can we distinguish meaningful structural changes from random 􀀍uctuations. Here we show that bootstrap resampling accompanied by signi􀀌cance clustering provides a solution to this problem. To connect changing structures with the changing function of networks, we highlight and summarize the signi􀀌cant structural changes with alluvial diagrams and realize de Solla Price's vision of mapping change in science: studying the citation pattern between about 7000 scienti􀀌c journals over the past decade, we 􀀌nd that neuroscience has transformed from an interdisciplinary specialty to a mature and stand-alone discipline. Introduction Researchers have developed a suite of network mapping tools to highlight important features while simplifying the overall structure of social and biological systems (1{6) . With such tools we can abstract, quantify, and comprehend the nature of systems with numerous and diverse interacting components. As powerful as these tools have proven to be for understanding a system's structure, we do not yet have an adequate tool for mapping how this structure changes. For example: How has the network of global air tra􀀎c changed over the past half century? How does the organization of social contacts change when diseases develop and spread? How does the network structure of the federal funds market change when credit markets freeze up? How do gene regulatory networks di􀀋er between cancer and non-cancer states? And how does science itself evolve as research tools, strategies, and agendas shift through time? To quantify change in large networks, we must 􀀌rst identify the important structures, then assess which of these structures are statistically signi􀀌cant, and 􀀌nally capture how these structures change. Any tool for analyzing change must distinguish between meaningful trends and statistical noise. For example, statistical network models and strati􀀌ed data make it possible to estimate global properties from the observation of sample networks (7{9) . But when we are interested in the unique identities of the individual network 􀀃 Electronic address: [rosvall@u.washington.edu](rosvall@u.washington.edu); URL: [http://www](http://www) . [tp.umu.se/](tp.umu.se/)~rosvall components | Chicago O'Hare plays a unique and irreplaceable role in the global air tra􀀎c network, for example | we need another approach. Recent network approaches have become prominent in the study of complex systems because they can capture and respect the identities and characteristics of the components (10, 11) . Often these individual di􀀋erences matter critically and clustering rather than strati􀀌cation must be employed to comprehend the data (1{6) . Moreover, many of the systems to which we apply network approaches are idiosyncratic in nature and preclude replicate observations. There is one and only one global air tra􀀎c network for the year 2009, for example. Therefore we cannot establish statistical signi􀀌cance by l

FAQs

Why is it difficult to map changes in large networks?

Because meaningful structural changes must be distinguished from statistical noise, and many systems do not allow replicate observations, leaving only a single observed network.

What method does the paper propose to quantify structural change?

It uses bootstrap resampling together with significance clustering to give statistical significance to partitions of a single network.

How are significant structural changes visualized and connected to function?

The paper highlights and summarizes significant structural changes with alluvial diagrams, linking structural evolution to changing network function.

Oct 26, 2026

Upload by Adam