BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Theoretical properties of Cook's Principal Fitted Components algor
 ithm - Oliver Johnson (University of Bristol)
DTSTART:20081031T160000Z
DTEND:20081031T173000Z
UID:TALK14338@talks.cam.ac.uk
CONTACT:8419
DESCRIPTION:I will talk about how to reduce the dimension of the linear re
 gression\nspace. In particular\, I will consider the theoretical propertie
 s of the\nestimators resulting from Dennis Cook's Principal Fitted Compone
 nts\nalgorithm. I will give sufficient conditions for root(n)-consistency 
 and\nexplain some of the simulation results in Cook's Fisher Lecture. I wi
 ll\nargue further that\, under Cook's model at least\, the PFC algorithm\n
 outperforms the more standard Principal Components algorithm.\n(paper to a
 ppear in Electronic Journal of Statistics)\n
LOCATION:MR12\, CMS\, Wilberforce Road\, Cambridge\, CB3 0WB
END:VEVENT
END:VCALENDAR