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 [R92].

Parameters

x : array_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

R92(1,2)

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