CategoricalIndicatorKriging

Indicator kriging of categories: the probability of each category kriged from its indicator, clipped to [0, 1] and rescaled to sum 1 at each target.

Parameters:

Name	Type	Description	Default
variograms	Variogram or sequence of Variogram	One indicator variogram per category, in code order, or a single one shared by all (one kriging system per target).	required
search	Search or sequence of Search	Neighborhood shared by all categories; a sequence is searched in passes, each filling the targets the previous left unestimated. With no ellipsoid, the middle category's variogram orients it.	required
simple	bool	Simple kriging with the declustered global proportions as means; ordinary kriging otherwise.	required
scheme	Categories	Names and colors of the categories: fit then takes labels, encoded by it. Without it, fit takes codes 0..k.	required

names property

names

Category names: those of scheme, else the codes as strings.

cross_validate method descriptor

cross_validate(*, folds=None)

Re-estimates each sample's probabilities from the others, through the same search passes, each in its own domain, against the proportions of all samples.

Parameters:

Name	Type	Description	Default
folds	int	Leave-one-out when None; otherwise k-fold, each sample estimated without the samples of its fold. Samples fitted with holes keep their holes whole: the j-th of the sorted hole ids goes to fold j % folds, an untagged sample i to fold i % folds.	None

Returns:

Type	Description
CategoricalCrossValidation

fit method descriptor

fit(
    coords,
    categories,
    *,
    weights=None,
    holes=None,
    domains=None,
    domain_column=None,
)

Stores the samples. Samples sharing a location keep the first one, with a warning naming their holes.

Parameters:

Name	Type	Description	Default
coords	(array_like, PointSet or BlockModel)		required
categories	array_like or str	A category per sample: a label of scheme, or without it a code 0..k; or the column of coords holding them. So for the other per-sample arguments.	required
weights	array_like or str	Declustering weights for the global proportions: the simple kriging means.	None
holes	array_like or str	Drill-hole ids or names, for max_per_hole.	None
domains	array_like	Domain label of each sample. A target is then estimated from the samples of its own domain, plus those of other domains within Search soft; predict then needs domains too.	None
domain_column	str	The column of coords holding the domains; instead of domains.	None

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,
    *,
    domains=None,
    domain_column=None,
    anisotropy=None,
    diagnostics=False,
)

Probability of each category at each target.

Parameters:

Name	Type	Description	Default
targets	(array_like, PointSet or BlockModel)		required
domains	array_like	Domain label of each target, or one label for all; a domain without samples is left unestimated.	None
domain_column	str	The column of targets holding the domains; instead of domains.	None
anisotropy	LocalAnisotropy	Orients every category's variogram and the search at each target.	None
diagnostics	bool	Fill CategoricalIndicatorSummary.diagnostics.	False

Returns:

Type	Description
CategoricalIndicatorSummary	NaN where the search found too few samples.

to_parquet method descriptor

to_parquet(path)

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

used in

• storing containers in parquet
• domain cleanup
• domain change tables
• categorical indicator kriging