statsmodels.stats.covariance.corr_quadrant#

statsmodels.stats.covariance.corr_quadrant(data, transform=<ufunc 'sign'>, normalize=False)[source]#

Quadrant correlation

Status: unverified, subject to change

Parameters:
dataarray_like

2-D data with observations in rows and variables in columns

transformcallable

Function used to transform the demeaned data before computing the correlation. Default is np.sign.

normalizebool

If True, normalize the resulting matrix by the standard deviations so that it is a proper correlation matrix. Default is False.

Returns:
corrndarray

correlation matrix

References

[1]

Croux, Christophe, and Catherine Dehon. “Influence Functions of the Spearman and Kendall Correlation Measures.” Statistical Methods & Applications 19, no. 4 (May 12, 2010): 497-515. https://doi.org/10.1007/s10260-010-0142-z.