BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:On Neighbourhood Cross Validation
DTSTART:20230714T151500
DTEND:20230714T170000
DTSTAMP:20260408T121759Z
UID:da6c395a59e3eb955b4362a10b485778236f09556d010f56fb0e158d
CATEGORIES:Conferences - Seminars
DESCRIPTION:Simon Woods\, School of Mathematics\, University of Edinburgh\
 nCross validation comes in many varieties\, but some of the more interesti
 ng flavours require multiple model fits with consequently high cost. This 
 talk shows how the high cost can be side-stepped for a wide range of model
 s estimated using a quadratically penalized smooth loss\, with rather low 
 approximation error.\n\nOnce the computational cost has the same leading o
 rder as a single model fit\, it becomes feasible to efficiently optimize t
 he chosen cross-validation criterion with respect to multiple smoothing/pr
 ecision parameters. Interesting applications include cross-validating smoo
 th additive quantile regression models\, and the use of leave-out-neighbou
 rhood cross validation for dealing with nuisance short range autocorrelati
 on.\n\nThe link between cross validation and the jackknife can be exploite
 d to  obtain reasonably well calibrated uncertainty quantification in the
 se cases.\n 
LOCATION:CM 1 221 https://plan.epfl.ch/?room==CM%201%20221
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR