module data.preprocessing.base_scaler

Base Scaler.

This module encapsulates functionality to scale, i.e. normalize, streaming observations. The abstract BaseScaler should be used to implement custom scaling methods. A scaler object can be provided to the data loader object.

Copyright (C) 2022 Johannes Haug.

class BaseScaler

Abstract Base Class for online data scaling.

Attributes:

  • reset_after_drift (bool): A boolean indicating if the scaler will be reset after a drift was detected.

method BaseScaler.__init__

__init__(reset_after_drift: bool)

Initializes the data scaler.

Args:

  • reset_after_drift: A boolean indicating if the scaler will be reset after a drift was detected.

method BaseScaler.partial_fit

partial_fit(
    X: Union[numpy._array_like._SupportsArray[numpy.dtype], numpy._nested_sequence._NestedSequence[numpy._array_like._SupportsArray[numpy.dtype]], bool, int, float, complex, str, bytes, numpy._nested_sequence._NestedSequence[Union[bool, int, float, complex, str, bytes]]]
)

Updates the scaler.

Args:

  • X: Array/matrix of observations.

method BaseScaler.reset

reset()

Resets the scaler.

method BaseScaler.transform

transform(
    X: Union[numpy._array_like._SupportsArray[numpy.dtype], numpy._nested_sequence._NestedSequence[numpy._array_like._SupportsArray[numpy.dtype]], bool, int, float, complex, str, bytes, numpy._nested_sequence._NestedSequence[Union[bool, int, float, complex, str, bytes]]]
) → Union[numpy._array_like._SupportsArray[numpy.dtype], numpy._nested_sequence._NestedSequence[numpy._array_like._SupportsArray[numpy.dtype]], bool, int, float, complex, str, bytes, numpy._nested_sequence._NestedSequence[Union[bool, int, float, complex, str, bytes]]]

Scales the given observations.

Args:

  • X: Array/matrix of observations.

Returns:

  • ArrayLike: The scaled observations.

This file was automatically generated via lazydocs.