Brenton M. Wiernik

I had forgotten to remove the zero-inflation bit. Thanks for checking. With respect to the values, the "more complex model" heuristic you are using just doesn't work outside of normal...

> I'm not sure I get why a more complex model wouldn't give a better fit here. R2 does not indicate "better fit" except in the case of linear regression....

alist() is base

We can extract the raw `rstan` object from an `"ulam"` class object with: ```r d

`ulam` objects have a `stanfit` slot. ```r str(m, max.level = 2) # Formal class 'ulam' [package "rethinking"] with 9 slots # ..@ call : language rethinking::ulam(flist = alist(x ~ dnorm(mu,...

The problem is that that there isn't one R2 for a model with a dispersion model. There's one for every possible value of sigma. We should perhaps consider treating these...

I can take care of improving the outliers plots

I have formulas for all of the SEM fit indices under the sun. I can take that.

Regarding out of sample stuff, I would also love if performance could do various true cross-validation stuff, both individual residuals and aggregates like RMSE, R^2^, and deviance. Those are currently...

Regarding PCA/factor analysis, to meet most user expectations, that's going to require at least a handful of dependencies like GPArotation. Various nuances of rotations and extraction methods are pretty fiddly,...