Likelihood-based assessment of dynamic networks
Likelihood-based assessment of dynamic networks
This paper deals with the problem of assessing probabilistic models that represent the evolution of a target graph. Such models have long been a topic of interest for a number of networks, especially communications networks. The solution developed in this paper gives a rigorous way to calculate the likelihood of the observed graph evolution having arisen from a wide variety of hypothesized models encompassing many already present in the literature. The framework is shown to recover parameters from artificial data and is tested on real data sets from Facebook and from emails from the company Enron.
1-18
Clegg, Richard
994a2792-68a5-4a96-8530-d4a09128b4ad
Parker, Ben
26c5a5ab-17b3-4d6c-ae11-abf3a2554529
Rio, Miguel
b6a22a93-a53c-4487-be23-55834358c539
Clegg, Richard
994a2792-68a5-4a96-8530-d4a09128b4ad
Parker, Ben
26c5a5ab-17b3-4d6c-ae11-abf3a2554529
Rio, Miguel
b6a22a93-a53c-4487-be23-55834358c539
Clegg, Richard, Parker, Ben and Rio, Miguel
(2016)
Likelihood-based assessment of dynamic networks.
Journal of Complex Networks, .
(doi:10.1093/comnet/cnv031).
Abstract
This paper deals with the problem of assessing probabilistic models that represent the evolution of a target graph. Such models have long been a topic of interest for a number of networks, especially communications networks. The solution developed in this paper gives a rigorous way to calculate the likelihood of the observed graph evolution having arisen from a wide variety of hypothesized models encompassing many already present in the literature. The framework is shown to recover parameters from artificial data and is tested on real data sets from Facebook and from emails from the company Enron.
Text
feta_comnet_2015.pdf
- Accepted Manuscript
More information
Accepted/In Press date: 14 December 2015
e-pub ahead of print date: 23 March 2016
Organisations:
Statistical Sciences Research Institute
Identifiers
Local EPrints ID: 397485
URI: http://eprints.soton.ac.uk/id/eprint/397485
ISSN: 2051-1310
PURE UUID: 2a71d69f-5531-41f2-b892-781b39dbc8ce
Catalogue record
Date deposited: 01 Jul 2016 14:04
Last modified: 15 Mar 2024 05:42
Export record
Altmetrics
Contributors
Author:
Richard Clegg
Author:
Ben Parker
Author:
Miguel Rio
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics