CategoricalIndicatorSummary

Probability of each category at every target from categorical indicator kriging; NaN where unestimated.

correction property

correction

Sum over categories of |corrected − kriged| probability.

diagnostics property

diagnostics

Per-target Table when predicted with diagnostics=True, else None: n_samples, pass (the search, from 1, that filled the target), n_holes (distinct holes among the samples used; untagged samples count one each), mean_distance (to the samples used), max_samples_reached (1 where the search returned max_samples), correction and n_order_violations (categories kriged outside [0, 1]). Null where unestimated.

entropy property

entropy

Entropy of the probabilities scaled to [0, 1]: 0 where one category is certain, 1 where all are equally likely.

most_likely property

most_likely

Code of the most probable category, ties to the lowest; float so that unestimated targets are NaN, as Categories.decode takes it.

names property

names

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

probabilities property

probabilities

(targets, categories) probabilities, each row summing to 1.

proportions property

proportions

Declustered proportion of each category among the samples.

from_parquet staticmethod

from_parquet(path)

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

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

• storing containers in parquet