Is it possible to fit a Weibull survival regression model with multivariate priors on its parameters? To keep things simple, assume there are no covariates, so only the shape and scale (or mean according to the Weibull parametrization) are estimated. Instead of using univariate priors for those two parameters, would it be possible and make sense to use a multivariate prior with LKJ for correlations in this case?
Hi
I guess you’ve seen this? If I were you I’d try it out with MV prior LKJ and report back :)
Hello! Did you mean to post a link? :)
Hehehe, I’m an idiot :)