BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Latent Mixture Quantile Regression for Longitudinal Data - Dr. Bo 
 Fu\, School of Community Based Medicine\, University of Manchester
DTSTART:20110628T133000Z
DTEND:20110628T143000Z
UID:TALK30834@talks.cam.ac.uk
CONTACT:Li Su
DESCRIPTION:\nAbstract\n\nI will talk about mixture median and quantile mo
 dels for describing latent\ngrowth curves of longitudinal outcomes. The mo
 dels exhibit the different latent\nclasses of evolution of the underlying 
 outcome process. The mixture median model\ncan be used as a robust alterna
 tive to the Gaussian likelihood based latent class\nmodel for skewed data\
 , and the quantile models provide a complete regression picture\nfor inves
 tigating the latent class structure at different quantiles. The within-sub
 ject\ncorrelation is incorporated by a marginal approach based on the idea
  of\nweighting. The weighted estimating equations for the model parameters
  are given\,\nand the asymptotic distribution of the resultant estimates i
 s established to approximate\nthe standard errors of the parameter estimat
 es. A penalized weighted loss\nfunction is defined to select the optimal n
 umber of latent classes. The proposed\nmethods are illustrated with data f
 rom 418 arthritis patients recruited between\n1990-1994.
LOCATION:Large Seminar Room\, 1st Floor\, Institute of Public Health\, Uni
 versity Forvie Site\, Robinson Way\, Cambridge
END:VEVENT
END:VCALENDAR