Generating orthogonal latent variables

6

I can give the latent vectors independent normal priors, but this doesn’t do what I’m looking for, which is that the vectors are defined in such a way that each individual realization of vector 1 is orthogonal to vector 2. It looks like this could be done by generating independent normals then applying a qr decomposition Random Orthogonal Matrices - #7 by simonbyrne - Optimization (Mathematical) - Julia Programming Language. I’ll give that a try, thanks