*Inserm Workshops
**
**Only a few days left to register to :*
*
_Inserm Workshop 205:_ Mixture Modelling for longitudinal data
**
Phase I , June 2-4, 2010 - Saint-Raphaël
**Phase II , June 7-8, 2010 - Paris*
*Registration deadline: _April 2, 2010_*
*For more information and registration, please visit the following link:
*Inserm Workshop 205
<http://www.rh.inserm.fr/INSERM/IntraRh/RHPublication.nsf/mDisplayMotsClefsWeb?OpenForm&arg1=17&arg2=>
*or email to : [log in to unmask]*
*Aims :*
Present, discuss and contrast several approaches to analyse heterogeneous
longitudinal data and highlight profiles of change with time. Both cluster
analysis methods and mixture models for longitudinal data will be presented
with applications to medical and psychological data. Particular emphasis
will
be put on parameter interpretation, post-fit evaluation of model assumptions
and conditions of use of the various approaches. The workshop will also
include a presentation of joint models with latent classes for time-to-event
and longitudinal outcome and an introduction to some other latent variable
models.
*Program*
The first session will focus on exploratory methods of clustering with the
presentation and application of PROC TRAJ and Kml tools. Then, after an
introduction to mixed models for longitudinal data, the mixture model for
longitudinal data will be presented and applied, in particular using Mplus
software. The third session will be dedicated to extension of these
models to the
analysis of multivariate longitudinal data, to the joint analysis of
longitudinal
data and time-to-event, and to non-parametric estimation. Finally, latent
variable models will be addressed with the presentation of a latent process
model for heterogeneous profiles of evolution, and the application of
structural
equation models using Mplus.
--------------------
Hélène Jacqmin-Gadda
Directrice de Recherche
Centre de Recherche INSERM U 897,
Equipe de Biostatistique
Bordeaux
France
Site : http://www.isped.u-bordeaux2.fr/
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|