BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Optimal dimension-reduced calibration and the terminal case for sp
 atial models - Daniel Williamson (University of Exeter)
DTSTART:20180307T114500Z
DTEND:20180307T123000Z
UID:TALK102262@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Since the seminal paper by Kennedy and O&rsquo\;Hagan in 2001\
 , the calibration of computer models using Gaussian process emulators has 
 represented a gold standard for scientists and statisticians working to qu
 antify uncertainty using complex computer codes. When the output of such c
 odes is high dimensional\, such as with the spatial fields routinely produ
 ced by climate models\, the standard approach (attributed to Higdon in 200
 8) is to take principal components across the model output\, and use the l
 oadings on these as a lower dimensional representation of the model output
  that can be used within the Kennedy and O&rsquo\;Hagan framework.   In th
 is talk I will argue that\, in general\, we should not expect this to work
 . I will introduce what we term a &ldquo\;terminal case analysis&rdquo\; f
 or general computer model calibration\, show the implications for inferenc
 e of a terminal case analysis and argue that though a high dimensional com
 puter model may not lead to a terminal case analysis\, the standard statis
 tical treatment outlined above invariably leads to one artificially. I wil
 l then present our solution to this which uses rotation ideas to fix the s
 earch directions of our lower dimensional representations so that general 
 calibration of spatio-temporal models is possible. We apply our method to 
 idealised examples and to the output of the state of the art Canadian atmo
 sphere model CanAGCM4. We will see that the problem of calibrating climate
  models requires a great deal of novel statistical thinking before we\, as
  a community\, can claim to have a solution ready for this important appli
 cation area. This is work done with (and by) Dr James Salter.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR