How can we build a mixture of mixtures?

Questions
5

Thanks for this hint… But I certainly misunderstood something since It makes an error.

I have two vectors of dimension 10 and 5, I’d like to stack them to obtain a 15 dimensions vector…
This is my code :

w = [ ]
w.append(pm.Dirichlet('w_1',a=np.array([0.0000001]*10)))
w.append(pm.Dirichlet('w_2',a=np.array([1]*5)))
w_ = tt.stacklists(w)

The error occurs on the stacklists line: 

"ValueError: all the input array dimensions except for the concatenation axis must match exactly"

So, just for a try (because it’s not what I’m looking for), I set the same dimension (10) for all the vectors :

w = [ ]
w.append(pm.Dirichlet(‘w_1’,a=np.array([0.0000001]*10)))
w.append(pm.Dirichlet(‘w_2’,a=np.array([1]*10)))
w_ = tt.stacklists(w)

Then The stack operation seems to work, at least it do not produce error… But it fails later in my code because the dimension of w_ is now 10 (and should be 15 for my application).
But this surprises me because I though that the dimension of w_ would be 20 then (10+10),
so I think I really don’t understand how to stack/group random variables…

Any help will be appreciated.