A likelihood approach to analysis of network data

Carsten Wiuf; Markus Brameier; Oskar Hagberg; Michael P.H. Stumpf

doi:10.1073/pnas.0600061103

A likelihood approach to analysis of network data

Carsten Wiuf^*, Markus Brameier, Oskar Hagberg, Michael P.H. Stumpf

^*Corresponding author for this work

51 Citations (Scopus)

Abstract

Biological, sociological, and technological network data are often analyzed by using simple summary statistics, such as the observed degree distribution, and nonparametric bootstrap procedures to provide an adequate null distribution for testing hypotheses about the network. In this article we present a full-likelihood approach that allows us to estimate parameters for general models of network growth that can be expressed in terms of recursion relations. To handle larger networks we have developed an importance sampling scheme that allows us to approximate the likelihood and draw inference about the network and how it has been generated, estimate the parameters in the model, and perform parametric bootstrap analysis of network data. We illustrate the power of this approach by estimating growth parameters for the Caenorhabditis elegans protein interaction network.

Original language	English
Journal	Proceedings of the National Academy of Sciences of the United States of America
Volume	103
Issue number	20
Pages (from-to)	7566-7570
Number of pages	5
ISSN	0027-8424
DOIs	https://doi.org/10.1073/pnas.0600061103
Publication status	Published - 16 May 2006
Externally published	Yes

Keywords

Biological network
Importance sampling
Likelihood recursion
Network model
Random graph

Access to Document

10.1073/pnas.0600061103

Cite this

@article{fb4e86043a024f1a80e241176ac03ef4,

title = "A likelihood approach to analysis of network data",

abstract = "Biological, sociological, and technological network data are often analyzed by using simple summary statistics, such as the observed degree distribution, and nonparametric bootstrap procedures to provide an adequate null distribution for testing hypotheses about the network. In this article we present a full-likelihood approach that allows us to estimate parameters for general models of network growth that can be expressed in terms of recursion relations. To handle larger networks we have developed an importance sampling scheme that allows us to approximate the likelihood and draw inference about the network and how it has been generated, estimate the parameters in the model, and perform parametric bootstrap analysis of network data. We illustrate the power of this approach by estimating growth parameters for the Caenorhabditis elegans protein interaction network.",

keywords = "Biological network, Importance sampling, Likelihood recursion, Network model, Random graph",

author = "Carsten Wiuf and Markus Brameier and Oskar Hagberg and Stumpf, {Michael P.H.}",

year = "2006",

month = may,

day = "16",

doi = "10.1073/pnas.0600061103",

language = "English",

volume = "103",

pages = "7566--7570",

journal = "Proceedings of the National Academy of Sciences of the United States of America",

issn = "0027-8424",

publisher = "The National Academy of Sciences of the United States of America",

number = "20",

}

TY - JOUR

T1 - A likelihood approach to analysis of network data

AU - Wiuf, Carsten

AU - Brameier, Markus

AU - Hagberg, Oskar

AU - Stumpf, Michael P.H.

PY - 2006/5/16

Y1 - 2006/5/16

N2 - Biological, sociological, and technological network data are often analyzed by using simple summary statistics, such as the observed degree distribution, and nonparametric bootstrap procedures to provide an adequate null distribution for testing hypotheses about the network. In this article we present a full-likelihood approach that allows us to estimate parameters for general models of network growth that can be expressed in terms of recursion relations. To handle larger networks we have developed an importance sampling scheme that allows us to approximate the likelihood and draw inference about the network and how it has been generated, estimate the parameters in the model, and perform parametric bootstrap analysis of network data. We illustrate the power of this approach by estimating growth parameters for the Caenorhabditis elegans protein interaction network.

AB - Biological, sociological, and technological network data are often analyzed by using simple summary statistics, such as the observed degree distribution, and nonparametric bootstrap procedures to provide an adequate null distribution for testing hypotheses about the network. In this article we present a full-likelihood approach that allows us to estimate parameters for general models of network growth that can be expressed in terms of recursion relations. To handle larger networks we have developed an importance sampling scheme that allows us to approximate the likelihood and draw inference about the network and how it has been generated, estimate the parameters in the model, and perform parametric bootstrap analysis of network data. We illustrate the power of this approach by estimating growth parameters for the Caenorhabditis elegans protein interaction network.

KW - Biological network

KW - Importance sampling

KW - Likelihood recursion

KW - Network model

KW - Random graph

UR - http://www.scopus.com/inward/record.url?scp=33646745354&partnerID=8YFLogxK

U2 - 10.1073/pnas.0600061103

DO - 10.1073/pnas.0600061103

M3 - Journal article

C2 - 16682633

AN - SCOPUS:33646745354

SN - 0027-8424

VL - 103

SP - 7566

EP - 7570

JO - Proceedings of the National Academy of Sciences of the United States of America

JF - Proceedings of the National Academy of Sciences of the United States of America

IS - 20

ER -

A likelihood approach to analysis of network data

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this