BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:Isaac Newton Institute Seminar Series
SUMMARY:Goodness of fit of logistic models for random grap
hs - Pierre Latouche (Université Paris 1)
DTSTART;TZID=Europe/London:20160726T093000
DTEND;TZID=Europe/London:20160726T100000
UID:TALK66847AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/66847
DESCRIPTION:We consider \;binary networks along with cova
riate information on the edges. In order to take t
hese covariates into account\, logistic-type model
s for random graphs are often considered. One of t
he main questions which arises in practice is to a
ssess the goodness of fit of a \;model. To ad
dress this problem\, we add a general term\, relat
ed to the graphon function of W-graph models\, to
the logistic models. Such an extra term can be app
roximated from a blockwise constant function obtai
ned using \;stochastic block models with incr
easing number of clusters. If the given network is
fully explained by the covariates\, then a sole b
lock should be estimated from data. This framework
allows to derive a testing procedure from a model
based selection context. Bayes factors or posteri
or odds can then be used for decision making. Over
all\, the logistic model considered necessitates t
wo types of variational approximations to derive t
he model selection approach. \;
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR