Module preprocessing 

Feature preprocessing transformers.

All transformers maintain per-feature online statistics and use O(d) memory where d is the feature dimension.

Structs

Clipper

A simple element-wise clipper that bounds feature values to a fixed range.

ConstantImputer

Replaces NaN values with a constant.

ConstantImputerConfig

Configuration for ConstantImputer.

ForwardFill

Replaces NaN values with the last observed non-NaN value.

FrequencyEncoder

Online frequency encoder for string features.

MeanImputer

Replaces NaN values with the per-feature running mean.

MinMaxScaler

Online min-max scaler that scales features to a configurable range.

MissingIndicator

Adds a missing-value indicator for each feature.

OneHotEncoder

Online one-hot encoder for string features.

OrdinalEncoder

Online ordinal encoder for string features.

StandardScaler

Online standard scaler that standardizes features to approximately zero mean and unit variance.

StandardScalerConfig

Configuration for StandardScaler.