BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Conditional Density Estimation through Enriched Dirichlet Process 
 Mixture Models - Sara Wade\, University of Cambridge
DTSTART:20130205T143000Z
DTEND:20130205T153000Z
UID:TALK42669@talks.cam.ac.uk
CONTACT:Dr Jack Bowden
DESCRIPTION:Flexible conditional density estimation can be achieved by mod
 elling the joint density of the response and covariate as a Dirichlet proc
 ess mixture. An appealing aspect of this approach is that computations are
  relatively easy. In this talk\, I will discuss the predictive performance
  of these models with an increasing number of covariates. Even for a moder
 ate number of covariates\, we find that the likelihood for x tends to domi
 nate the posterior of the latent random partition\, degrading the predicti
 ve performance of the model. To overcome this\, we suggest using a differe
 nt nonparametric prior\, namely an Enriched Dirichlet process. Our proposa
 l maintains a simple allocation rule\, so that computations remain relativ
 ely simple. Advantages will be shown through both predictive equations and
  examples\, including an application to diagnosis Alzheimer's disease. 
LOCATION:Large  Seminar Room\, 1st Floor\, Institute of Public Health\, Un
 iversity Forvie Site\, Robinson Way\, Cambridge
END:VEVENT
END:VCALENDAR