pywddff

Python package to perform feature engineering using wavelet transforms, geared towards those who work with tabular data and machine learning models to forecast time series.

Documentation

You can find the documentation for pywddff including API references here.

About the name

pywddff stands for "Python Wavelet Data Driven Forecasting Framework". WDDFF is an approach to combining wavelet transform generated features and machine learning for forecasting time series. The WDDFF paper can be found here.

Installation

This package is not yet listed on the Python Package Index (PyPI). You can use poetry install to install pywddff by first cloning this repository and then running poetry install from the root directory. Please note that you must have poetry already installed and available from your terminal. See here for instructions on installing poetry on your system.

$ git clone https://github.com/johnswyou/pywddff.git
$ cd pywddff
$ poetry install

Update (June 6, 2023): You can now pip install

$ pip install pywddff

Available wavelet/scaling filters

pywddff gives access to 128 (decomposition level 1) wavelet (high pass) and scaling (low pass) filters.

The following are short names for the available filters:

['bl7', 'bl9', 'bl10', 
'beyl', 
'coif1', 'coif2', 'coif3', 'coif4', 'coif5', 
'db1', 'db2', 'db3', 'db4', 'db5', 'db6', 'db7', 'db8', 'db9', 'db10', 'db11', 'db12', 
'db13', 'db14', 'db15', 'db16', 'db17', 'db18', 'db19', 'db20', 'db21', 'db22', 'db23', 
'db24', 'db25', 'db26', 'db27', 'db28', 'db29', 'db30', 'db31', 'db32', 'db33', 
'db34', 'db35', 'db36', 'db37', 'db38', 'db39', 'db40', 'db41', 'db42', 'db43', 'db44', 'db45', 
'fk4', 'fk6', 'fk8', 'fk14', 'fk18', 'fk22', 
'han2_3', 'han3_3', 'han4_5', 'han5_5', 
'dmey', 
'mb4_2', 'mb8_2', 'mb8_3', 'mb8_4', 'mb10_3', 'mb12_3', 'mb14_3', 'mb16_3', 'mb18_3', 'mb24_3', 'mb32_3', 
'sym2', 'sym3', 'sym4', 'sym5', 'sym6', 'sym7', 'sym8', 'sym9', 'sym10', 'sym11', 'sym12', 'sym13', 'sym14', 
'sym15', 'sym16', 'sym17', 'sym18', 'sym19', 'sym20', 'sym21', 'sym22', 'sym23', 'sym24', 'sym25', 'sym26', 'sym27', 
'sym28', 'sym29', 'sym30', 'sym31', 'sym32', 'sym33', 'sym34', 'sym35', 'sym36', 'sym37', 'sym38', 'sym39', 'sym40', 
'sym41', 'sym42', 'sym43', 'sym44', 'sym45', 
'vaid', 
'la8', 'la10', 'la12', 'la14', 'la16', 'la18', 'la20']

where

• bl corresponds to "Best-localized Daubechies"
• beyl corresponds to "Beylkin"
• coif corresponds to "Coiflets"
• db corresponds to "Daubechies"
• fk corresponds to "Fejér-Korovkin"
• han corresponds to "Han linear-phase moments"
• dmey corresponds to "Discrete Meyer"
• mb corresponds to "Morris minimum-bandwidth"
• sym corresponds to "Symlets"
• vaid corresponds to "Vaidyanathan"
• la corresponds to "Least asymmetric"

NOTE: The db1 is the Haar wavelet.

Source of data

The wfilters function from MATALB R2022b's wavelet toolbox provided 121 of the 128 wavelet (scaling) filters available. The R package wavelets provided the remaining 7 least asymmetric wavelet (scaling) filters.

Contributing

Interested in contributing? Check out the contributing guidelines. Please note that this project is released with a Code of Conduct. By contributing to this project, you agree to abide by its terms.

License

pywddff was created by John You. It is licensed under the terms of the MIT license.

Credits

pywddff was created with cookiecutter and the py-pkgs-cookiecutter template.