Are extreme value estimation methods useful for network data?

Phyllis Wan-Huen, Tiandong Wang, Richard A. Davis, Sidney I. Resnick

Research output: Contribution to journalArticle

Abstract

Preferential attachment is an appealing edge generating mechanism for modeling social networks. It provides both an intuitive description of network growth and an explanation for the observed power laws in degree distributions. However, there are often difficulties fitting parametric network models to data due to either model error or data corruption. In this paper, we consider semi-parametric estimation based on an extreme value approach that begins by estimating tail indices of the power laws of in- and out-degree for the nodes of the network using nodes with large in- and out-degree. This method uses tail behavior of both the marginal and joint degree distributions. We compare the extreme value method with the existing parametric approaches and demonstrate how it can provide more robust estimates of parameters associated with the network when the data are corrupted or when the model is misspecified.

Original languageEnglish (US)
JournalExtremes
DOIs
StateAccepted/In press - Jan 1 2019

Fingerprint

Extreme Values
Degree Distribution
Power Law
Tail Index
Robust Estimate
Preferential Attachment
Semiparametric Estimation
Tail Behavior
Model Error
Vertex of a graph
Parametric Model
Joint Distribution
Social Networks
Network Model
Intuitive
Modeling
Demonstrate
Extreme values
Node
Power law

Keywords

  • Estimation
  • Multivariate heavy-tailed statistics
  • Power laws
  • Preferential attachment
  • Regular variation

ASJC Scopus subject areas

  • Statistics and Probability
  • Engineering (miscellaneous)
  • Economics, Econometrics and Finance (miscellaneous)

Cite this

Are extreme value estimation methods useful for network data? / Wan-Huen, Phyllis; Wang, Tiandong; Davis, Richard A.; Resnick, Sidney I.

In: Extremes, 01.01.2019.

Research output: Contribution to journalArticle

Wan-Huen, Phyllis ; Wang, Tiandong ; Davis, Richard A. ; Resnick, Sidney I. / Are extreme value estimation methods useful for network data?. In: Extremes. 2019.
@article{a4effb70d99042b1bbac625a7080dd2f,
title = "Are extreme value estimation methods useful for network data?",
abstract = "Preferential attachment is an appealing edge generating mechanism for modeling social networks. It provides both an intuitive description of network growth and an explanation for the observed power laws in degree distributions. However, there are often difficulties fitting parametric network models to data due to either model error or data corruption. In this paper, we consider semi-parametric estimation based on an extreme value approach that begins by estimating tail indices of the power laws of in- and out-degree for the nodes of the network using nodes with large in- and out-degree. This method uses tail behavior of both the marginal and joint degree distributions. We compare the extreme value method with the existing parametric approaches and demonstrate how it can provide more robust estimates of parameters associated with the network when the data are corrupted or when the model is misspecified.",
keywords = "Estimation, Multivariate heavy-tailed statistics, Power laws, Preferential attachment, Regular variation",
author = "Phyllis Wan-Huen and Tiandong Wang and Davis, {Richard A.} and Resnick, {Sidney I.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/s10687-019-00359-x",
language = "English (US)",
journal = "Extremes",
issn = "1386-1999",
publisher = "Springer Netherlands",

}

TY - JOUR

T1 - Are extreme value estimation methods useful for network data?

AU - Wan-Huen, Phyllis

AU - Wang, Tiandong

AU - Davis, Richard A.

AU - Resnick, Sidney I.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Preferential attachment is an appealing edge generating mechanism for modeling social networks. It provides both an intuitive description of network growth and an explanation for the observed power laws in degree distributions. However, there are often difficulties fitting parametric network models to data due to either model error or data corruption. In this paper, we consider semi-parametric estimation based on an extreme value approach that begins by estimating tail indices of the power laws of in- and out-degree for the nodes of the network using nodes with large in- and out-degree. This method uses tail behavior of both the marginal and joint degree distributions. We compare the extreme value method with the existing parametric approaches and demonstrate how it can provide more robust estimates of parameters associated with the network when the data are corrupted or when the model is misspecified.

AB - Preferential attachment is an appealing edge generating mechanism for modeling social networks. It provides both an intuitive description of network growth and an explanation for the observed power laws in degree distributions. However, there are often difficulties fitting parametric network models to data due to either model error or data corruption. In this paper, we consider semi-parametric estimation based on an extreme value approach that begins by estimating tail indices of the power laws of in- and out-degree for the nodes of the network using nodes with large in- and out-degree. This method uses tail behavior of both the marginal and joint degree distributions. We compare the extreme value method with the existing parametric approaches and demonstrate how it can provide more robust estimates of parameters associated with the network when the data are corrupted or when the model is misspecified.

KW - Estimation

KW - Multivariate heavy-tailed statistics

KW - Power laws

KW - Preferential attachment

KW - Regular variation

UR - http://www.scopus.com/inward/record.url?scp=85070249540&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070249540&partnerID=8YFLogxK

U2 - 10.1007/s10687-019-00359-x

DO - 10.1007/s10687-019-00359-x

M3 - Article

JO - Extremes

JF - Extremes

SN - 1386-1999

ER -