Unable to specify a correct prior distribution

Estimation bayesian
8

I think the only way to do that would be to specify each of the diagonal elements separately, i.e. something like

@random begin
  ηCL  ~ InverseGamma(...)
  ηV1  ~ InverseGamma(...)
  ηV2  ~ InverseGamma(...)
end

I chose InverseGamma here since it’s equivalent to a scalar InverseWishart but it might be better to use the Gamma since it has support at zero.

1 Like