pyanno4rt.learning_model.preprocessing.transformers

Data transformers module.

The module aims to provide methods and classes for data transformation.

Overview

Classes
`Identity`	Identity transformer class.
`StandardScaler`	Standard scaling transformer class.
`Whitening`	Whitening transformer class.

Attributes
`transformer_map`	-

Classes

class pyanno4rt.learning_model.preprocessing.transformers.Identity[source]

Identity transformer class.

This class provides methods to fit, transform and gradientize the input features by their identity (default preprocessing step).

Overview

Methods
`fit`(features, labels)	Fit the identity transformer.
`transform`(features, labels)	Transform the input features/labels.
`compute_gradient`(features)	Compute the identity transformer gradient w.r.t the input features.

Members

fit(features, labels)[source]

Fit the identity transformer.

Parameters:

features (ndarray) – Values of the input features.
labels (None or ndarray) – Values of the input labels.

transform(features, labels)[source]

Transform the input features/labels.

Parameters:

features (ndarray) – Values of the input features.
labels (None or ndarray) – Values of the input labels.

Returns:

ndarray – Transformed values of the input features.
None or ndarray – Transformed values of the input labels.

compute_gradient(features)[source]

Compute the identity transformer gradient w.r.t the input features.

Parameters:: features (ndarray) – Values of the input features.
Returns:: Value of the identity transformer gradient.
Return type:: ndarray

class pyanno4rt.learning_model.preprocessing.transformers.StandardScaler(center=True, scale=True)[source]

Standard scaling transformer class.

This class provides methods to fit, transform and gradientize the input features by their z-score.

Parameters:

center (bool, default=True) – Indicator for the centering of the data by the mean values.
scale (bool, default=True) – Indicator for the scaling of the data by the standard deviations.

center

See ‘Parameters’.

Type:: bool

scale

See ‘Parameters’.

Type:: bool

means

Mean values of the features (if center is false, set to zeros).

Type:: ndarray

deviations

Standard deviations of the features (if scale is false, set to ones).

Type:: ndarray

Overview

Methods
`fit`(features, labels)	Fit the standard scaling transformer.
`transform`(features, labels)	Transform the input features/labels.
`compute_gradient`(features)	Compute the standard scaling transformer gradient w.r.t the input features.

Members

fit(features, labels)[source]

Fit the standard scaling transformer.

Parameters:

features (ndarray) – Values of the input features.
labels (None or ndarray) – Values of the input labels.

transform(features, labels)[source]

Transform the input features/labels.

Parameters:

features (ndarray) – Values of the input features.
labels (None or ndarray) – Values of the input labels.

Returns:

ndarray – Transformed values of the input features.
None or ndarray – Transformed values of the input labels.

compute_gradient(features)[source]

Compute the standard scaling transformer gradient w.r.t the input features.

Parameters:: features (ndarray) – Values of the input features.
Returns:: Value of the standard scaling transformer gradient.
Return type:: ndarray

class pyanno4rt.learning_model.preprocessing.transformers.Whitening(method='zca')[source]

Whitening transformer class.

This class provides methods to fit, transform and gradientize the input features by their whitening matrix.

Parameters:

method ({'pca', 'zca'}, default='zca') –

Method for the computation of the whitening matrix.

’zca’ : zero-phase component analysis (Mahalanobis transformation)
’pca’ : principal component analysis

method

See ‘Parameters’.

Type:: {‘pca’, ‘zca’}

means

Mean values of the features.

Type:: ndarray

matrix

Whitening matrix.

Type:: ndarray

Overview

Methods
`fit`(features, labels)	Fit the whitening transformer.
`transform`(features, labels)	Transform the input features/labels.
`compute_gradient`(features)	Compute the whitening transformer gradient w.r.t the input features.

Members

fit(features, labels)[source]

Fit the whitening transformer.

Parameters:

features (ndarray) – Values of the input features.
labels (None or ndarray) – Values of the input labels.

transform(features, labels)[source]

Transform the input features/labels.

Parameters:

features (ndarray) – Values of the input features.
labels (None or ndarray) – Values of the input labels.

Returns:

ndarray – Transformed values of the input features.
None or ndarray – Transformed values of the input labels.

compute_gradient(features)[source]

Compute the whitening transformer gradient w.r.t the input features.

Parameters:: features (ndarray) – Values of the input features.
Returns:: Value of the whitening transformer gradient.
Return type:: ndarray

Attributes

pyanno4rt.learning_model.preprocessing.transformers.transformer_map