statsmodels.stats.dist_dependence_measures.distance_variance

statsmodels.stats.dist_dependence_measures.distance_variance(x)[source]

Distance variance.

Calculate the empirical distance variance as described in [1].

Parameters
xarray_like, 1-D or 2-D

If x is 1-D than it is assumed to be a vector of observations of a single random variable. If x is 2-D than the rows should be observations and the columns are treated as the components of a random vector, i.e., each column represents a different component of the random vector x.

Returns
float

The empirical distance variance of x.

References

1

Szekely, G.J., Rizzo, M.L., and Bakirov, N.K. (2007) “Measuring and testing dependence by correlation of distances”. Annals of Statistics, Vol. 35 No. 6, pp. 2769-2794.

Examples

>>> from statsmodels.stats.dist_dependence_measures import
... distance_variance
>>> distance_variance(np.random.random(1000))
0.21732609190659702