BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Local and discrete scoring rules - Matt Parry (Cambridge)
DTSTART:20080314T140000Z
DTEND:20080314T150000Z
UID:TALK10282@talks.cam.ac.uk
CONTACT:8048
DESCRIPTION:At its simplest\, a scoring rule is a loss function for choosi
 ng a certain distribution to represent the uncertainty of a random variabl
 e. (A scoring rule is not a score function!) One of the key requirements i
 s honesty: the expected score must be minmised by choosing the true distri
 bution --- if it is known. I take an introductory tour of the role of scor
 ing rules in decision theory and statistical inference and of their connec
 tion to entropy and geometry. I end with a discussion of recent advances i
 n the use of scoring rules in cases of both continuous and discrete sample
  spaces. \n
LOCATION:MR12\, CMS\, Wilberforce Road\, Cambridge\, CB3 0WB
END:VEVENT
END:VCALENDAR