setu.Standardizer

class setu.Standardizer(mean, std)

Bases: Module

Z-scoring transform for data standardization.

Transforms data to have zero mean and unit standard deviation. Use Standardizer.fit() to create from data.

mean

Mean values per dimension, shape (dim,).

std

Standard deviation per dimension, shape (dim,). Clipped to minimum 1e-10 to avoid division by zero.

Parameters:

• mean (Array)

• std (Array)

classmethod fit(data)

Fit standardizer to data.

Parameters:

data (Array) – Input data (n_samples, dim)

Return type:

Standardizer

Returns:

Fitted Standardizer

transform(data)

Transform data to standardized space.

Parameters:

data (Array) – Input data (…, dim)

Return type:

Array

Returns:

Standardized data with same shape

inverse(data)

Transform standardized data back to original space.

Parameters:

data (Array) – Standardized data (…, dim)

Return type:

Array

Returns:

Original-scale data with same shape