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:Latent Space Stochastic Block Model for Social Net
works - Brendan Murphy (University College Dublin)
DTSTART;TZID=Europe/London:20160726T113000
DTEND;TZID=Europe/London:20160726T120000
UID:TALK66849AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/66849
DESCRIPTION:A large number of statistical models have been pro
posed for social network analysis in recent years.
In this paper\, we propose a new model\, the
\;latent position stochastic block model\, which e
xtends and generalises both latent space model (Ho
ff et al.\, 2002) and stochastic block model (Nowi
cki and Snijders\, 2001). The probability of an ed
ge between two actors in a network depends on thei
r respective class labels as well as latent positi
ons in an unobserved latent space. The proposed mo
del is capable of representing transitivity\, clus
tering\, as well as disassortative mixing. A Bayes
ian method with Markov chain Monte Carlo sampling
is proposed for estimation of model parameters. Mo
del selection is performed by directly estimating
marginal likelihood for each model and models of d
ifferent number of classes or dimensions of latent
space can be compared. We apply the network model
to one simulated network and two real social netw
orks.
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR