Cokriging

Cokriging of variable from samples of several variables under a linear model of coregionalization; means switches to simple cokriging.

fit method descriptor

fit(coords, values, variables, *, holes=None)

variables gives each sample's variable index. Samples of one variable sharing a location keep the first, with a warning naming their holes. values, variables and holes may name columns of coords.

from_parquet staticmethod

from_parquet(path)

Reads to_parquet output; raises InvalidInput on another class's file or a newer format.

predict method descriptor

predict(
    targets,
    *,
    variable=0,
    return_variance=False,
    anisotropy=None,
    collocated=None,
)

Estimates of variable; collocated maps a variable index to its values at every target, or to the column of targets holding them, for collocated cokriging. anisotropy (a LocalAnisotropy) gives each target its own search ellipsoid and the anisotropy of every structure and cross term, taken from the nearest location.

to_parquet method descriptor

to_parquet(path)

Writes arrays as Parquet columns and parameters as JSON in the file metadata; from_parquet reads it back.

used in

• intrinsic coregionalization
• cokriging