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:Designing efficient composite likelihoods - Cristi
ano Varin (UniversitĂ CĂ Foscari di Venezia)
DTSTART;TZID=Europe/London:20170704T141500
DTEND;TZID=Europe/London:20170704T150000
UID:TALK73143AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/73143
DESCRIPTION:Composite likelihood is an inference function cons
tructed by compounding component likelihoods based
on low dimensional marginal or conditional distri
butions. Since the components are multiplied as if
they were independent\, the composite likelihood
inherits the properties of likelihood inference fr
om a misspecified model. The virtue of composite l
ikelihood inference is &ldquo\;combining the advan
tages of likelihood with computational feasibility
&rdquo\; (Reid\, 2013). Given the wide applicabili
ty\, composite likelihoods are attracting interest
as scalable surrogate for intractable likelihoods
. Despite the promise\, application of composite l
ikelihood is still limited by some theoretical and
computational issues which have received only par
tial or initial responses. Open theoretical questi
ons concern characterization of general model cond
itions assuring validity of composite likelihood i
nference\, optimal selection of component likeliho
ods and precise evaluation of estimation uncertain
ty. Computational issues concern how to design com
posite likelihood methods to balance statistical e
fficiency and computational efficiency.
In
this talk\, after a critical review of composite l
ikelihood theory\, I shall focus on the potential
merits of composite likelihood inference in modeli
ng temporal and spatial variation of disease incid
ence. The talk is based on past work with Nancy Re
id (Toronto) and David Firth (Warwick)\, and vario
us new \;projects with Manuela Cattelan (Padov
a)\, Xanthi Pedeli (Venice) and Guido Masarotto (P
adova).  \;
LOCATION:Seminar Room 1\, Newton Institute
CONTACT:INI IT
END:VEVENT
END:VCALENDAR