New functions for Exploratory Factor Analysis - Bartlett Sphericity Test and KMO test

[Rui Sarmento]

Dear contributers,

I'm planning to submit two new functions to the statsmodel module. Nonetheless, I'm not sure they are welcome so I would kindly ask if this is accepted. Additionally, this is the first time I submit to a Python repository so if you feel this might be a good contribution please guide me through the commit and pull request.

The two functions I wish to contribute are related to Exploratory Factor Analysis and the tests to evaluate if the sample is adequate to perform EFA.

Best Regards,

Rui

[josef...@gmail.com]

those are within the scope of statsmodels.

Everything with code, documentation and issues is done through github.
https://github.com/statsmodels/statsmodels/blob/master/CONTRIBUTING.rst
is our overview page for contributions and there are more details in
the developer docs.
Additional questions or issues can be discussed on mailing list or
github issues/PRs.
You can also check existing pull requests to see how it works.

One question is whether these functions should be in `multivariate` or
in `stats`.
`statsmodels.multivariate` currently only contains PCA but should be
extended to other multivariate methods.
Hypothesis test, descriptive statistics, confidence intervals are in
statsmodels.stats when they are not directly associated with a model.
e,g, various tests for the structure of correlation or covariance
matrices or two-sample comparisons could go into a
statsmodels.stats.covariance or correlation module. (that doesn't
exist yet)


Josef


>
> Best Regards,
>
> Rui