BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Fitting Latent Variable Models for Paired Comparisons and Ranking 
 Studies - An Application of Optimal Design Theory - Torsney\, B (Universit
 y of Glasgow)
DTSTART:20110901T090000Z
DTEND:20110901T093000Z
UID:TALK32611@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:In a paired comparisons experiment a subject has to indicate w
 hich of two 'treatments' Ti\, Tj is preferred. We observe Oij\, the freque
 ncy with which Ti is preferred to Tj.in nij comparisons. Under a class of 
 models for such data\, which include the Bradley Terry and Thurstone model
 s\, P(Ti is preferred to Tj) = F( i -  j)\, where F(.) is a symmetric dist
 ribution function and ( i)  is a treatment index. For identifiability   pu
 rposes constraints must be imposed on parameters. One is to assume that  i
 pi =  1\, where pi = ln( i)\; an alternative is  ipi =  1. Thus theorems i
 dentifying optimal design weights and algorithms for determining them carr
 y over to the maximum likelihood estimation of these parameters. \nOf cour
 se these tools can also be used  to determine locally optimal designs for 
 such models. \nWe will explore this fusion of topics\, taking the opportun
 ity to expand on the class of models\, both for simple  paired comparisons
  data and also for data consisting of orderings or rankings. In particular
  we will exploit multiplicative algorithms for maximum likelihood estimati
 on. \n\n\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR