Back to models
Transformer

WaveletPacketTransformer

API referenceSource
MultivariateUnequal-length series

Wavelet Packet Decomposition transformer.

Unlike the standard DWT which only decomposes the approximation branch at each level, wavelet packet decomposition recursively decomposes both approximation and detail branches, giving 2**level terminal sub-bands. This provides a richer frequency resolution.

Currently uses Haar wavelet coefficients internally.

Quickstart

python
from sktime.transformations.wavelet_packet import WaveletPacketTransformer

estimator = WaveletPacketTransformer(level=2, output_feature='energy')

Tags

Capabilities

  • Multivariate
  • Unequal-length series
  • Inverse transform: Not supported
  • Missing values: Not supported
  • Removes missing values: Not supported
  • Equalizes series length: Not supported

Properties

Input typescitype:transform-input
Series
Output typescitype:transform-output
Series
Label typescitype:transform-labels
None
Fit is emptyfit_is_empty
Yes
Keeps the time indextransform-returns-same-time-index
No
Requires Xrequires_X
Yes
Requires yrequires_y
No
X and y need the same indexX-y-must-have-same-index
No

Parameters(2)

levelint, default=2

Number of decomposition levels. Produces 2**level terminal sub-band nodes.

output_featurestr, default=”energy”

What to extract from each sub-band. One of:

  • "energy": sum of squared coefficients per node

  • "entropy": Shannon entropy of normalized coefficient power

  • "coefficients": concatenated raw packet coefficients

Examples

>>> from sktime.transformations.wavelet_packet import (
... WaveletPacketTransformer,
... )
>>> from sktime.datasets import load_airline
>>> y = load_airline ()
>>> t = WaveletPacketTransformer (level = 2, output_feature = "energy")
>>> y_features = t. fit_transform (y)