sgt regression

[texmiller]

Thanks for the great resources in this repo! I was wondering if it's possible to modify the likelihood functions for your univariate distributions so that parameters are vectors instead of reals, for regression applications. For example, I would like to build on your sgt example to estimate the parameters as linear functions of covariates. Thanks for any thoughts or tips.

[spinkney]

It should be fairly easy to add a vector signature. Did you want to try and create a pr?

[texmiller]

OK will give it a shot - thanks!