BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Goodness of fit of logistic models for random graphs - Pierre Lato
 uche (Université Paris 1)
DTSTART:20160726T083000Z
DTEND:20160726T090000Z
UID:TALK66847@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:We consider &nbsp\;binary networks along with covariate inform
 ation on the edges. In order to take these covariates into account\, logis
 tic-type models for random graphs are often considered. One of the main qu
 estions which arises in practice is to assess the goodness of fit of a &nb
 sp\;model. To address this problem\, we add a general term\, related to th
 e graphon function of W-graph models\, to the logistic models. Such an ext
 ra term can be approximated from a blockwise constant function obtained us
 ing &nbsp\;stochastic block models with increasing number of clusters. If 
 the given network is fully explained by the covariates\, then a sole block
  should be estimated from data. This framework allows to derive a testing 
 procedure from a model based selection context. Bayes factors or posterior
  odds can then be used for decision making. Overall\, the logistic model c
 onsidered necessitates two types of variational approximations to derive t
 he model selection approach.&nbsp\;
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR