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:Statistics
SUMMARY:A theoretical and methodological discussion of nes
ted sampling - Nicolas Chopin (Bristol)
DTSTART;TZID=Europe/London:20080118T140000
DTEND;TZID=Europe/London:20080118T150000
UID:TALK9975AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/9975
DESCRIPTION:Nested sampling is a novel simulation method for a
pproximating marginal likelihoods\, proposed by \\
cite{Skilling:2007a\,Skilling:2007b}. We establish
that nested sampling leads to an error that vanis
hes at the standard Monte Carlo rate N-1/2\, where
N is a tuning parameter that is proportional to t
he computational effort\, and that this error is a
symptotically Gaussian. We show that the correspon
ding asymptotic variance typically grows linearly
with the dimension of the parameter. We use these
results to discuss the applicability and efficien
cy of nested sampling in realistic problems\, incl
uding posterior distributions for mixtures. We pro
pose an extension of nested sampling that makes it
possible to avoid resorting to MCMC to obtain the
simulated points. We study two alternative method
s for computing marginal likelihood\, which\, in c
ontrast with nested sampling\, are based on draws
from the posterior distribution and we conduct a c
omparison with nested sampling on several realisti
c examples. \n
LOCATION:MR12\, CMS\, Wilberforce Road\, Cambridge\, CB3 0W
B
CONTACT:
END:VEVENT
END:VCALENDAR