I did not understand the dependent example you gave, moreover in the labeling of second diagram when arrows were from Y to X, the name was given XY. But for arrow from X to Z, the name was XZ instead of ZX? The picture is confusing me because of this. And it would have been better if an even simpler example was used, as I did not get the dependent scenario.

Thanks! and Cheers. ]]>

thanks for providing your code. I'm in the process of porting your code to PyMC3, but get quite different results for the robust model. Unfortunately, I could not find out why this happens. I'd appreciate if you could have a look at http://stackoverflow.com/questions/40126809/robust-bayesian-correlation-with-pymc3, maybe you have an idea.

Thanks!

]]>thanks for your fantastic work, this is very helpful! I am not very familiar with PyMC yet, how could one give two independent priors on mu and sigma? I.e. instead of

mu = Normal('mu', 0, 0.000001, size=2)

sigma = Uniform('sigma', 0, 1000, size=2)

I would like to give mu1, mu2, sigma1, sigma2, but then don't know to combine them into one mu and one sigma. ]]>