RobustScaler#

class mvpy.preprocessing.RobustScaler(with_centering: bool = True, with_scaling: bool = True, quantile_range: tuple[float, float] = (25.0, 75.0), dims: list | tuple | int | None = None)[source]#

Implements a robust scaler that is invariant to outliers.

By default, this scaler removes the median before scaling the data according to the interquartile range \([0.25, 0.75]\). This is useful because, unlike Scaler, it means that RobustScaler is robust to outliers that might affect a Scaler poorly.

Both centering and scaling are optional and can be turned on or off using with_centering and with_scaling.

Parameters:

with_centeringbool, default=True: If True, center the data before scaling.
with_scalingbool, default=True: If True, scale the data according to the quantiles.
quantile_rangetuple[float, float], default=(25.0, 75.0): Tuple describing the quantiles.
dimsint, list or tuple of ints, default=None: The dimensions over which to scale (None for first dimension).

Attributes:

with_centeringbool, default=True: If True, center the data before scaling.
with_scalingbool, default=True: If True, scale the data according to the quantiles.
quantile_rangetuple[float, float], default=(25.0, 75.0): Tuple describing the quantiles.
dimsint, list or tuple of ints, default=None: The dimensions over which to scale (None for first dimension).
dims_tuple[int], default=None: Tuple specifying the dimensions to scale over.
centre_torch.Tensor, default=None: The centre of each feature of shape X.
scale_torch.Tensor, default=None: The scale of each feature of shape ``X`.

See also

mvpy.preprocessing.Scaler: An alternative scaler that normalises data to zero mean and unit variance.
mvpy.preprocessing.Clamp: A complementary class that implements clamping data at specific values.

Examples

>>> import torch
>>> from mvpy.preprocessing import RobustScaler
>>> scaler = RobustScaler().to_torch()
>>> X = torch.normal(5, 10, (1000, 5))
>>> X[500,0] = 1e3
>>> X.std(0)
tensor([32.9122,  9.9615, 10.1481, 10.1058,  9.7468])
>>> Z = scaler.fit_transform(X)
>>> Z.std(0)
tensor([2.7348, 0.7351, 0.7464, 0.7609, 0.7154])
>>> H = scaler.inverse_transform(Z)
>>> H.std(0)
tensor([32.9122,  9.9615, 10.1481, 10.1058,  9.7468])

clone() → RobustScaler[source]#

Obtain a clone of this class.

Returns:

scalerRobustScaler: The cloned robust scaler.

copy() → RobustScaler[source]#

Obtain a copy of this class.

Returns:

scalerRobustScaler: The copied robust scaler.

fit(X: ndarray | Tensor, *args: Any) → RobustScaler[source]#

Fit the scaler.

Parameters:

Xnp.ndarray | torch.Tensor: The data of arbitrary shape.
argsAny: Additional arguments.

Returns:

scalersklearn.base.BaseEstimator: The fitted scaler.

fit_transform(X: ndarray | Tensor, *args: Any) → ndarray | Tensor[source]#

Fit and transform the data in one step.

Parameters:

Xnp.ndarray | torch.Tensor: The data of shape X.
argsAny: Additional arguments.

Returns:

Znp.ndarray | torch.Tensor: The transformed data of shape X.

inverse_transform(X: ndarray | Tensor, *args: Any) → ndarray | Tensor[source]#

Invert the transform of the data.

Parameters:

Xnp.ndarray | torch.Tensor: The data of shape X.
argsAny: Additional arguments.

Returns:

Xnp.ndarray | torch.Tensor: The inverse transformed data of shape X.

to_numpy() → _RobustScaler_numpy[source]#

Select the numpy backend. Note that this cannot be called for conversion.

Returns:

scaler_RobustScaler_numpy: The robust scaler using the numpy backend.

to_torch() → _RobustScaler_torch[source]#

Select the torch backend. Note that this cannot be called for conversion.

Returns:

scaler_RobustScaler_torch: The robust scaler using the torch backend.

transform(X: ndarray | Tensor, *args: Any) → ndarray | Tensor[source]#

Transform the data using scaler.

Parameters:

Xnp.ndarray | torch.Tensor: The data of shape X.
argsAny: Additional arguments.

Returns:

Znp.ndarray | torch.Tensor: The transformed data of shape X.

RobustScaler#

This Page