Skip to content

API Reference

PreprocessorConfig

PreprocessorConfig dataclass

Configuration for data preprocessors.

Attributes:

Name Type Description
name Literal

Name of the preprocessor.
categorical_name Literal

Name of the categorical encoding method. Valid options are "none", "numeric", "onehot", "ordinal", "ordinal_shuffled". Default is "none".
append_original bool

Whether to append the original features to the transformed features. Default is False.
subsample_features float

Fraction of features to subsample. -1 means no subsampling. Default is -1.
global_transformer_name str

Name of the global transformer to use. Default is None.

ClassificationOptimizationMetricType

ClassificationOptimizationMetricType module-attribute 

ClassificationOptimizationMetricType = Literal[
    "auroc",
    "roc",
    "auroc_ovo",
    "balanced_acc",
    "acc",
    "log_loss",
    None,
]

RegressionOptimizationMetricType

RegressionOptimizationMetricType module-attribute 

RegressionOptimizationMetricType = Literal[
    "mse",
    "rmse",
    "mae",
    "r2",
    "mean",
    "median",
    "mode",
    "exact_match",
    None,
]