On the impact of network size and average degree on the robustness of centrality measures

Research output: Journal contributionsJournal articlesResearchpeer-review

Standard

On the impact of network size and average degree on the robustness of centrality measures. / Martin, Christoph; Niemeyer, Peter.
In: Network Science, Vol. 9, No. S1, 20.10.2021, p. S61-S82.

Research output: Journal contributionsJournal articlesResearchpeer-review

Harvard

APA

Vancouver

Bibtex

@article{9ad2ea4fef6849f5a5d19a4b14cdb56b,
title = "On the impact of network size and average degree on the robustness of centrality measures",
abstract = "Measurement errors are omnipresent in network data. Most studies observe an erroneous network instead of the desired error-free network. It is well known that such errors can have a severe impact on network metrics, especially on centrality measures: a central node in the observed network might be less central in the underlying, error-free network. The robustness is a common concept to measure these effects. Studies have shown that the robustness primarily depends on the centrality measure, the type of error (e.g., missing edges or missing nodes), and the network topology (e.g., tree-like, core-periphery). Previous findings regarding the influence of network size on the robustness are, however, inconclusive. We present empirical evidence and analytical arguments indicating that there exist arbitrary large robust and non-robust networks and that the average degree is well suited to explain the robustness. We demonstrate that networks with a higher average degree are often more robust. For the degree centrality and Erd{\AA}'s-R{\'e}nyi (ER) graphs, we present explicit formulas for the computation of the robustness, mainly based on the joint distribution of node degrees and degree changes which allow us to analyze the robustness for ER graphs with a constant average degree or increasing average degree.",
keywords = "Business informatics, centrality measures, measurement error, missing data, robustness",
author = "Christoph Martin and Peter Niemeyer",
note = "Special Issue S1: Complex Networks 2019. {\textcopyright} The Author(s), 2020. Published by Cambridge University Press ; International Conference on Complex Networks and their Applications - 2019 : Complex Networks ; Conference date: 10-12-2019 Through 12-12-2019",
year = "2021",
month = oct,
day = "20",
doi = "10.1017/nws.2020.37",
language = "English",
volume = "9",
pages = "S61--S82",
journal = "Network Science",
issn = "2050-1242",
publisher = "Cambridge University Press",
number = "S1",
url = "https://www.complexnetworks.org/index",

}

RIS

TY - JOUR

T1 - On the impact of network size and average degree on the robustness of centrality measures

AU - Martin, Christoph

AU - Niemeyer, Peter

N1 - Conference code: 8

PY - 2021/10/20

Y1 - 2021/10/20

N2 - Measurement errors are omnipresent in network data. Most studies observe an erroneous network instead of the desired error-free network. It is well known that such errors can have a severe impact on network metrics, especially on centrality measures: a central node in the observed network might be less central in the underlying, error-free network. The robustness is a common concept to measure these effects. Studies have shown that the robustness primarily depends on the centrality measure, the type of error (e.g., missing edges or missing nodes), and the network topology (e.g., tree-like, core-periphery). Previous findings regarding the influence of network size on the robustness are, however, inconclusive. We present empirical evidence and analytical arguments indicating that there exist arbitrary large robust and non-robust networks and that the average degree is well suited to explain the robustness. We demonstrate that networks with a higher average degree are often more robust. For the degree centrality and ErdÅ's-Rényi (ER) graphs, we present explicit formulas for the computation of the robustness, mainly based on the joint distribution of node degrees and degree changes which allow us to analyze the robustness for ER graphs with a constant average degree or increasing average degree.

AB - Measurement errors are omnipresent in network data. Most studies observe an erroneous network instead of the desired error-free network. It is well known that such errors can have a severe impact on network metrics, especially on centrality measures: a central node in the observed network might be less central in the underlying, error-free network. The robustness is a common concept to measure these effects. Studies have shown that the robustness primarily depends on the centrality measure, the type of error (e.g., missing edges or missing nodes), and the network topology (e.g., tree-like, core-periphery). Previous findings regarding the influence of network size on the robustness are, however, inconclusive. We present empirical evidence and analytical arguments indicating that there exist arbitrary large robust and non-robust networks and that the average degree is well suited to explain the robustness. We demonstrate that networks with a higher average degree are often more robust. For the degree centrality and ErdÅ's-Rényi (ER) graphs, we present explicit formulas for the computation of the robustness, mainly based on the joint distribution of node degrees and degree changes which allow us to analyze the robustness for ER graphs with a constant average degree or increasing average degree.

KW - Business informatics

KW - centrality measures

KW - measurement error

KW - missing data

KW - robustness

UR - http://www.scopus.com/inward/record.url?scp=85114751266&partnerID=8YFLogxK

U2 - 10.1017/nws.2020.37

DO - 10.1017/nws.2020.37

M3 - Journal articles

VL - 9

SP - S61-S82

JO - Network Science

JF - Network Science

SN - 2050-1242

IS - S1

T2 - International Conference on Complex Networks and their Applications - 2019

Y2 - 10 December 2019 through 12 December 2019

ER -

Documents

DOI

Recently viewed

Publications

  1. "Now I like it".
  2. Remotely sensed effectiveness assessments of protected areas lack a common framework
  3. Effects of preschoolers' storybook exposure and literacy environments on lower level and higher level language skills
  4. Review of transit data sources
  5. Incremental sheet forming with active medium
  6. Generalizing Trust
  7. The five-factor asset pricing model – A theoretical review and assessment
  8. Focus: Computational history and philosophy of science
  9. The self-sabotage of conservation
  10. Effects of Y Additions on the Microstructures and Mechanical Behaviours of as Cast Mg–xY–0.5Zr Alloys
  11. Politics after Networks
  12. Die Unternehmergesellschaft
  13. Pragmatic Function of Twitter Handlers' Perspectives on Children Discourse in Nigeria.
  14. Recycling-oriented fabrication of soft robots
  15. What do people do when they use the internet?
  16. New methods for the analysis of links between international firm activities and firm performance
  17. Introduction: A strategy for overcoming the definitional struggle
  18. Conceptual and procedural mathematical knowledge of beginning mathematics majors and preservice teachers
  19. Communication
  20. Reprint of: Drivers of within-tree leaf trait variation in a tropical planted forest varying in tree species richness
  21. The Timing of Daily Demand for Goods and Services
  22. The role of human resource practices for including persons with disabilities in the workforce
  23. The Balanced Scorecard and different Business Models in the textile industry
  24. Basin efficiency approach and its effect on streamflow quality, Zerafshan River Uzbekistan
  25. Logistical futures the chinese dream, debordering labor, and migration
  26. Appetizers for Business Integration into the heavy Meal of Transdisciplinary Practices
  27. Efficacy of cognitive bias modification interventions in anxiety and depression
  28. Differences in adaptation to light and temperature extremes of Chlorella sorokiniana strains isolated from a wastewater lagoon
  29. Culture as an Engine of Local Development Processes
  30. Quality Education and lifelong learning for all: trying to get to grips with the iridescent, multifaced, and at the same time universal character of SDG 4
  31. What do employers pay for employees’ complex problem solving skills?
  32. Exploring the uncanny valley effect in social robotics
  33. Dangerous settings and risky international assignments
  34. Building capacity for the science-policy interface on biodiversity and ecosystem services