I can talk about some results….that I might be able to add to the paper. I have this sticky issue of the theta’s moving to the edges of my simplex, yet the other papers I’ve read have results where the points do lie on the edges of the simplex. So I’m assuming they’ve encountered those issues as well.
Here are some results for the monk data. The simplex doesn’t look exactly like the Airoldi paper, but they used different data. An email was sent, but I think actually the plots are confirmation enough.
Another option for the paper is fit a simulated model and plot the logposterior for a single network?
And now for the Hierarchical Stuff:
Int Sim Networks Fitted Pretty
The Observation stuff is worse. But somehow also not worse. The predicted networks look similar – and the estimates for Beta and Tau are really accurate…except that I used pretty strong priors. I’m backing off that now and I trying to use more diffuse priors on the second run. But the question is, if I use more diffuse priors – won’t I kind of have an identifiability issue with tau and beta?
For example, gamma’s are being estimated as ~Gamma(10, 50) but with only 20 gammas, they could have come from ~Gamma(1,5) distribution. So should I somehow include this dependency in my MCMC? Not sure yet what to do.
These results are from with the following priors:
beta ~ G(50,1) and tau~G(10, 1) (I’m running (5, .1) and (1, .1) now.) And note that the data were generated using gamma~G(10,50). So I could try using a more diffuse prior to generate the data as well.
ObservationalSimsNetworksPretty