[SciPy-Dev] Fwd: Adding Biweight Mid-Correlation to scipy.stats

[Christopher McFarland]

Hi,

I have implemented a common "robust" correlation metric, the Biweight
Mid-Correlation [1], for myself and wonder if this would be welcomed
function for scipy.stats. This function in R is titled 'bicor' [2] and is
especially popular in the biology community.

If the addition is welcomed, I'd be happy to Unit Test/Document/properly
stylize the function. I also wonder if a unified correlation function (with
keyword argument method={'pearson', 'spearman', 'kendall', 'bicor'}) would
be welcomed. All these functions have a similar purpose, inputs, and
outputs, and are implemented in Pandas with simply one unified function.

Chris

[1] "Introduction to Robust Estimation and Hypothesis Testing", Rand
Wilcox, Academic Press, 1997.
[2] http://www.inside-r.org/packages/cran/WGCNA/docs/bicor
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scipy-dev/attachments/20150514/65ffe8fe/attachment.html>