Generalized Likelihood Maximization / Score Functions

[patrickeganfoley]

I'm writing some code for a special case of generalized likelihood maximization. To do this, I have functions

score(noise::Distribution, theta, observations)

and

information(noise::Distribution, theta, observations)

which implement the first and second derivative of a distribution's loglikelihood w/r/t its parameters. I use these so I can run a version of Fisher Scoring on several models. Are there plans to include first and second derivatives in some part of Stats in Julia?

[johnmyleswhite]

This could be cool to do.