Seminar : : Bayesian nonparametric analysis for surveys with randomly |
Other events...
|
Bayesian nonparametric analysis with the multinomial distribution and non-informative Dirichlet prior has an important application in survey analysis.
It provides a robust representation of the covariate distribution for incomplete covariates with randomly missing data. This allows fully efficient posterior analysis for generalized regression models by MCMC, and provides an alternative to multiple imputation, without the requirement of a parametric distribution for the incomplete covariates or the analysis of multiply imputed completed data sets.
Examples are given of 1- and 2-variable normal regression models.
Speaker(s) |
Murray and Irit Aitken
|
Location |
Mathematics Lecture Room 3 (G.02, Ground floor of Mathematics
|
|
Contact |
Annie Walker
<[email protected]>
: 3383
|
Start |
Thu, 28 Nov 2013 14:00
|
End |
Thu, 28 Nov 2013 15:00
|
Submitted by |
annie Walker <[email protected]>
|
Last Updated |
Tue, 26 Nov 2013 09:15
|
Included in the following Calendars: |
|
- Locations of venues on the Crawley and Nedlands campuses are
available via the Campus Maps website.
- Download this event as:
Text |
iCalendar
-
Mail this event:
|