Variational Bayesian inference for the Latent Position Cluster Model for network data
|Title:||Variational Bayesian inference for the Latent Position Cluster Model for network data||Authors:||Salter-Townshend, Michael; Murphy, Thomas Brendan||Permanent link:||http://hdl.handle.net/10197/12538||Date:||Jan-2013||Online since:||2021-10-11T14:29:28Z||Abstract:||A number of recent approaches to modeling social networks have focussed on embedding the nodes in a latent “social space”. Nodes that are in close proximity are more likely to form links than those who are distant. This naturally accounts for reciprocal and transitive relationships which are commonly found in many network datasets. The Latent Position Cluster Model is one such model that also explicitly incorporates clustering by modeling the locations using a finite Gaussian mixture model. Observed covariates and sociality random effects may also be modeled. However, inference for the model via MCMC is cumbersome and thus scaling to large networks is a challenge. Variational Bayesian methods offer an alternative inference methodology for this problem. Sampling based MCMC is replaced by an optimization that requires many orders of magnitude fewer iterations to converge. A Variational Bayesian algorithm for the Latent Position Cluster Model is therefore developed and demonstrated.||Funding Details:||Science Foundation Ireland||Type of material:||Journal Article||Publisher:||Elsevier||Journal:||Computational Statistics & Data Analysis||Volume:||57||Issue:||1||Start page:||661||End page:||671||Copyright (published version):||2012 Elsevier||Keywords:||Social network analysis; Variational Bayes; Latent position cluster model||DOI:||10.1016/j.csda.2012.08.004||Language:||en||Status of Item:||Peer reviewed||This item is made available under a Creative Commons License:||https://creativecommons.org/licenses/by-nc-nd/3.0/ie/|
|Appears in Collections:||Mathematics and Statistics Research Collection|
CASL Research Collection
Show full item record
If you are a publisher or author and have copyright concerns for any item, please email firstname.lastname@example.org and the item will be withdrawn immediately. The author or person responsible for depositing the article will be contacted within one business day.