A problem on Gibbs Sampler
- From: "hi" <jinyinglu@xxxxxxxxxxx>
- Date: 23 Sep 2006 05:08:59 -0700
I am trying to fit the ordinal data with this so-called location-shif
cummultive odds model
link(Pr(Yi<j))=(ai-bX)/exp(cX)
where the link can be probit or logit. I am trying to estimate the
coefficient via gibbs samppler. Howerver, I do not know how to comput
the full conditional distributions from the above model.
If any one know how to comput the full conditional distribution for
this model, I would greately appriate.
Jinying
.
- Prev by Date: "Seven out of every ten women preferred Patrick"
- Next by Date: ANN: Beta test invitation
- Previous by thread: "Seven out of every ten women preferred Patrick"
- Next by thread: ANN: Beta test invitation
- Index(es):