[SciPy-User] scipy.stats.models.mixed examples ?

[Skipper Seabold]

On Sun, Sep 20, 2009 at 7:50 AM,  <josef.pktd at gmail.com> wrote:
> Has anyone used scipy.stats.models.mixed and has some usage examples
> for it? An archive search came up empty.
>
> I wanted to see if it can be updated to integrate back into
> statsmodels from the sandbox. But I have a hard time seeing how it
> works (especially because of the internal use of formula). There is
> only a brief (incomplete) example in the file and not a single test
> for it. The results look reasonable and the implementation seems to
> follow the reference pretty closely. During gsoc, there wasn't enough
> time to clean it up and verify it.
>
> class Mixed(object):
>
>    """
>    Model for
>    EM implementation of (repeated measures)
>    mixed effects model.
>
>    \'Maximum Likelihood Computations with Repeated Measures:
>    Application of the EM Algorithm\'
>
>    Nan Laird; Nicholas Lange; Daniel Stram
>
>    Journal of the American Statistical Association,
>    Vol. 82, No. 397. (Mar., 1987), pp. 97-105.
>    """
>
> Skimming through the article it could be useful for some applications.
> Although for panel data, as far as I have seen, it wouldn't allow for
> time random effects, since there is no correlation between units
> allowed.
>
> Thanks,
>
> Josef

Not what you asked for, but a few references that I collected, as I
found the R stuff to be the most helpful in crafting examples to get
started.

http://stat.ethz.ch/R-manual/R-patched/library/nlme/html/lme.html

http://cran.r-project.org/doc/contrib/Fox-Companion/appendix-mixed-models.pdf

I have found the Kuhnert and Venables a good hands on intro for a
number of models as well:

http://cran.r-project.org/other-docs.html

Skipper