Adding Linear Mixed Model (LMM) support to SciRuby

[Frankenstein]

It's couple of hours before the deadline and I just realized I need a submission of code for pre-approval. Is it too late to submit code?

[Pjotr Prins]

Not too late.

On Fri, Mar 27, 2015 at 09:56:14AM -0700, Frankenstein wrote:
> It's couple of hours before the deadline and I just realized I need a
> submission of code for pre-approval. Is it too late to submit code?
>

> --
> You received this message because you are subscribed to the Google Groups
> "SciRuby Development" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [1]sciruby-dev...@googlegroups.com.
> For more options, visit [2]https://groups.google.com/d/optout.
>
> References
>
> Visible links
> 1. mailto:sciruby-dev...@googlegroups.com
> 2. https://groups.google.com/d/optout

[Frankenstein]

Can I have a suggested issue to work on?

[Frankenstein]

I'm looking up SciRuby Issue Trackers now. A little guide will be great.

[Claudio Bustos]

About (G)LMM, to begin we should:
- Create a little document with the theory behind it, and estimation
methods (marginal quasi-likehood, penalized quasi-likehood, MCMC,
Gaussian Quadrature...)
- Using some of the named methods, estimate the parameters using
matrix for fixed and mixed effects.
- Calculate point estimates and variances for all the mixed effects
- Design a little parser to transform formulas on (design) matrix, a la R.



--
Claudio Bustos
Psicólogo
clbu...@gmail.com