Is there a way to separately specify conjugate priors for the variance and correlations of a multivariate normal? The inverse Wishart is conjugate if you want to specify the covariance, but covariance matrices aren't very intuitive. Is it possible to instead specify a prior over the correlations, and separately a prior over the marginal variance for each parameter, while still keeping conjugacy?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.