fips.covariance.CorrelationModel#

class fips.covariance.CorrelationModel[source]#

Bases: ABC

A correlation model: how correlated two labels are, from their coordinates.

The base of Exponential, Gaussian, Matern, Constant, Independent and Kronecker. A model has no labels of its own: with_std() scales it into a covariance, which reads the model’s levels from the labels of the space it is put on.

A new model is a frozen dataclass that subclasses this one, names the levels it reads in levels and implements correlation().

See also

fips.covariance.CovarianceModel

A model with its standard deviations.

Attributes

levels

The levels of the labels the correlation is a function of.

Methods

correlation(coords)

Return the correlation between the rows of coords.

with_std(std)

Return the covariance of this correlation with standard deviations std.