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:Non-parametric regression with observations at imp
recise times: Bayesian radiocarbon calibration - T
im Heaton (Sheffield)
DTSTART;TZID=Europe/London:20110603T160000
DTEND;TZID=Europe/London:20110603T170000
UID:TALK30799AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/30799
DESCRIPTION:For accurate radiocarbon dating\, it is necessary
to identify fluctuations in\nthe level of radioact
ive carbon (14C) over time. The processes that giv
e\nrise to these fluctuations are not understood a
nd so a datum-based\ncalibration curve is required
. Historic material can provide us with\nobservati
ons of radiocarbon levels at distinct times in the
past\, but in\nmany cases the precise age of this
material is unknown and must itself be\nestimated
.\n\nWith some exceptions\, the current non-parame
tric regression literature\nconcentrates on the ca
se where the times at which we observe the functio
n\nare known precisely. In this talk we will prese
nt a general methodology for\nnon-parametric regre
ssion\, developed as part of this on-going radioca
rbon\ncalibration work\, which can handle such imp
recise time estimates.\n\nIn our approach\, the un
derlying function is modelled as a Weiner process\
nwith sample paths that are updated\, in light of
the imprecise data\, through\nuse of a Metropolis-
within-Gibbs algorithm. This methodology is\ncompu
tationally efficient due to the independent increm
ent property of the\nprocess and use of a Brownian
bridge.\n\n\nhttp://maths.dept.shef.ac.uk/pas/sta
ff_info.php?id=289
LOCATION:MR12\, CMS\, Wilberforce Road\, Cambridge\, CB3 0W
B
CONTACT:Richard Nickl
END:VEVENT
END:VCALENDAR