classify

classify(
    criteria,
    rules,
    *,
    default="unclassified",
    domains=None,
    domain_column=None,
) -> ndarray

Labels each block with the first rule whose conditions all hold.

Parameters:

Name	Type	Description	Default
criteria	Table or dict of str to array_like	Per-block metrics, e.g. slope, efficiency from predict(..., diagnostics=True), nearest_dist, n_holes from neighborhood_stats and distances from hole_distance.	required
rules	sequence of (str, dict), or dict of domain to such a sequence	(label, {metric: (op, threshold)}) in priority order; op is one of <, <=, >, >=. NaN never satisfies a condition. With domains, a dict gives each domain its own rules; the None key covers domains not listed, and blocks of other domains get default.	required
default	str	Label where no rule holds.	'unclassified'
domains	array_like	Domain of each block, the keys of rules.	None
domain_column	str	The column of criteria holding the domains; instead of domains.	None

Examples:

>>> classify(d, [("measured", {"slope": (">=", 0.8), "n_holes": (">=", 3)}),
...              ("indicated", {"slope": (">=", 0.5)})], default="inferred")

used in

• Checking a model
• classification
• A coal seam in a lease
• From drill holes to a classified model