dcase_models.data.Scaler

class dcase_models.data.Scaler(normalizer='standard')

Bases: object

Scaler object to normalize or scale the data.

Parameters:

normalizer : {‘standard’ or ‘minmax’}, default=’standard’

Type of normalizer.

See also

DataGenerator

data generator class

Examples

>>> from dcase_models.data.scaler import Scaler
>>> import numpy as np
>>> scaler = Scaler('minmax')
>>> X = 3 * np.random.rand(10, 150)
>>> print(np.amin(X), np.amax(X))

>>> scaler.fit(X)
>>> X = scaler.transform(X)
>>> print(np.amin(X), np.amax(X))

Attributes:

scaler : sklearn.preprocessing.StandardScaler or list

Scaler object for standard normalizer or list for minmax scaler.

__init__(normalizer='standard')

Initialize the Scaler.

If normalizer is ‘standard’, initialize the sklearn object.

Methods

__init__([normalizer])	Initialize the Scaler.
fit(X[, inputs])	Fit the Scaler.
inverse_transform(X)	Invert transformation.
partial_fit(X)	Fit the Scaler in one batch.
transform(X)	Scale X using the scaler.

fit(X, inputs=True)

Fit the Scaler.

Parameters:

X : ndarray or DataGenerator

Data to be used in the fitting process.

inverse_transform(X)

Invert transformation.

Parameters:

X : ndarray

Data to be scaled.

Returns:

ndarray

Scaled data. The shape of the output is the same of the input.

partial_fit(X)

Fit the Scaler in one batch.

Parameters:

X : ndarray

Data to be used in the fitting process.

transform(X)

Scale X using the scaler.

Parameters:

X : ndarray

Data to be scaled.

Returns:

ndarray

Scaled data. The shape of the output is the same of the input.