mvdist_bivariate_empirical

mvdist_bivariate_empirical(
    x1,
    x2,
    p,
    x1_transform='None',
    x2_transform='None',
    p_transform='None',
)

Construct a bivariate empirical distribution.

Mirrors the C# BivariateEmpirical class of the Numerics library: a joint distribution defined over a grid of two marginal value vectors and a matrix of associated probabilities.

Parameters

Name	Type	Description	Default
x1	array-like of float	Grid values for each dimension.	required
x2	array-like of float	Grid values for each dimension.	required
p	array - like	len(x1) x len(x2) matrix of joint probabilities.	required
x1_transform	('None', 'Logarithmic', 'NormalZ')	How each axis is interpolated between.	"None"
x2_transform	('None', 'Logarithmic', 'NormalZ')	How each axis is interpolated between.	"None"
p_transform	('None', 'Logarithmic', 'NormalZ')	How each axis is interpolated between.	"None"

Returns

Name	Type	Description
	MultivariateDistribution	Family "BivariateEmpirical". Note: pdf is an upstream stub (see :meth:MultivariateDistribution.pdf).

Examples

>>> mv = mvdist_bivariate_empirical([1, 2], [1, 2], [[0.2, 0.3], [0.2, 0.3]])
>>> mv.cdf([1.5, 1.5]) >= 0
True