BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Stein's Paradox - Dr Richard Samworth (DPMMS)
DTSTART:20120224T190000Z
DTEND:20120224T200000Z
UID:TALK36116@talks.cam.ac.uk
CONTACT:25292
DESCRIPTION:(Free for members\, £2 for non-members)\n\nStein's paradox is
  one of the most surprising results in Statistics. Suppose X1\,...\,Xp are
  independent random variables\, with Xi ~ N(θi\,1). If we want to estimat
 e θ = (θ1\,...\,θp)\, the most obvious choice is to use X = (X1\,...\,X
 p). It turns out that\, provided p>2\, we can find a better estimator\, in
  a very natural sense that I will make precise. As well as giving the (fai
 rly straightforward) proof\, I will discuss geometric intuition and other 
 explanations for this result\, and discuss extensions. I will also show ho
 w the improved estimator can be used to give good predictions of baseball 
 batting averages.
LOCATION:Centre for Mathematical Sciences (CMS)
END:VEVENT
END:VCALENDAR