What is the proper syntax for specifying theta terms in a multivariate mixture model? I’d like to be able to estimate mixing parameters that are constrained to be equal (ie. theta2 for sepal length is the same as theta2 for sepal width), but would settle for just setting fixed values too.
Example:
data(iris) iris.mvmix<-brm(cbind(Sepal.Length,Sepal.Width)~1, family=mixture(gaussian,nmix=2,order=T), data=iris)
You currently can’t fix parameters across univariate models in a multivariate models with brms.
This will be a feature of brms 3.0 that is still at least a few months away.
For the time being, you have to write the Stan code yourself, perhaps starting with the
output of make_stancode.
No worries, just wanted to make sure I wasn’t missing a bit of syntax before I switched over to base STAN. Thanks!
Was wondering if this constraint is possible for the newer version of brms? I know we are not still in version 3
I am able to make the changes to the make_stancode and followong the steps in the thread I can update the brms opbject.
Thanks for the work on brms
It is unfortunately not yet possible in brms natively.
I imagined so, appreciate all the feautures from brms