Standardize

class sdm.processing.numerical.Standardize(*, eps: float = 0.0)

Bases: Processor, InvertibleMixin

Center and scale each feature column.

Constant columns use a unit scale to keep the transform finite and invertible. NaN and infinite values are ignored when fitting statistics and preserved during the transform.

Parameters:

eps (float) – Value added to each fitted standard deviation.

with_state(mean: Tensor | float, scale: Tensor | float) → Self

Set fixed statistics in place and disable refitting.

Parameters:

• mean (Tensor | float) – The mean to subtract.

• scale (Tensor | float) – The scale to divide by.

Return type:

Self

Capabilities

Handled Semantic Types	numerical
Requires Fitting	✅