PySDKit 0.4.23

In this merger, we've added a new signal decomposition algorithm, Feature Mode Decomposition. This algorithm is primarily used to analyze faulty equipment, targeting mechanical fault characteristics—i.e., the signal's impulse and periodicity—as decomposition targets.

Thanks to @ for contributing the Python code.

What's Changed

• Feature Mode Decomposed algorithm to be implemented. by @JacktheFowler in #44
• Update PySDKit 0.4.23 for FMD by @wwhenxuan in #45

New Contributors

• @JacktheFowler made their first contribution in #44

Full Changelog: 0.4.22...0.4.23

PySDKit 0.4.22

We have updated the writing of union data types in Python. All the | that appeared in Python 3.10 have been updated to the old version of Union, making it applicable to versions 3.8 and 3.9.

What's Changed

• Using correct fft function in VME by @feng1m8 in #36
• Preliminary completion of STL algorithm by @changewam in #37
• Delete .idea directory by @wwhenxuan in #38
• Master by @wwhenxuan in #40
• Master by @wwhenxuan in #43

New Contributors

• @feng1m8 made their first contribution in #36

Full Changelog: 0.4.20...0.4.22

PySDKit 0.4.21

What's Changed

• Using correct fft function in VME by @feng1m8 in #36
• Preliminary completion of STL algorithm by @changewam in #37
• Delete .idea directory by @wwhenxuan in #38
• Update PySDKit 0.4.21 for STL by @wwhenxuan in #39

New Contributors

• @feng1m8 made their first contribution in #36

Full Changelog: 0.4.20...0.4.21

---

@changewam have added STL, a new signal decomposition algorithm, which can effectively decompose time series data into trends, cycles, and noise through optimization algorithms.

We would like to thank @feng1m8 for helping us fix the error in the VME algorithm regarding Fourier transform.

PySDKit 0.4.20

What's Changed

• Update the Generation for Univariate Signal Test Data by @wwhenxuan in #34
• Update the Version of PySDKit 0.4.20 by @wwhenxuan in #35

Full Changelog: 0.4.19...0.4.20

PySDKit 0.4.19

What's Changed

• Add New Signal Decomposition Algorithms Robust Local Mean Decomposition by @wwhenxuan in #32
• Fix EFD: complex array by @josefinez in #31
• Update Some Information for PySDKit 0.4.19 by @wwhenxuan in #33

New Contributors

• @josefinez made their first contribution in #31

Full Changelog: 0.4.18...0.4.19

PySDKit 0.4.18

Installation

Install via pip and dive into our [demo](https://github.com/wwhenxuan/PySDKit/blob/main/example/demo.ipynb) to kickstart your experience—quick and easy! 🥰

pip install pysdkit

What's Changed

• Update HHT by @wwhenxuan in #27
• Update Version by @wwhenxuan in #28

Full Changelog: 0.4.17...0.4.18

PySDKit 0.4.17

Install via pip and dive into our [demo](https://github.com/wwhenxuan/PySDKit/blob/main/example/demo.ipynb) to kickstart your experience—quick and easy! 🥰

pip install pysdkit

We added new algorithms: Empirical Fourier Decomposition 🥳

What's Changed

• Init NFMD by @wwhenxuan in #24
• Update the new signal decompsition algorithm: EFD by @wwhenxuan in #25
• Update New Version by @wwhenxuan in #26

Full Changelog: 0.4.16...0.4.17

PySDKit 0.4.16

Install via pip and dive into our [demo](https://github.com/wwhenxuan/PySDKit/blob/main/example/demo.ipynb) to kickstart your experience—quick and easy! 🥰

pip install pysdkit

What's Changed

• Update faemd.py by @wwhenxuan in #7
• Update faemd.py for FAEMD by @wwhenxuan in #8
• Update FA-EMD and testing by @wwhenxuan in #9
• FIX wrong for saving IMFs in FA-EMD by @wwhenxuan in #10
• Master by @wwhenxuan in #11
• Update packages version by @wwhenxuan in #12
• Add KNN for classification by @wwhenxuan in #13
• Update for init.py by @wwhenxuan in #14
• try by @changewam in #16
• First Commit for Doc by @wwhenxuan in #17
• Master by @wwhenxuan in #18
• Update init.py for more information by @wwhenxuan in #19
• Update the JMD for signal decomposition by @wwhenxuan in #20
• Update VME by @wwhenxuan in #21
• Update EMD2D by @wwhenxuan in #22
• Update Version by @wwhenxuan in #23

New Contributors

• @wwhenxuan made their first contribution in #7
• @changewam made their first contribution in #16

Full Changelog: 0.4.15...0.4.16

PySDKit 0.4.15

Exciting News! 🎉 PySDKit's first pre-release version is now available!

Install via pip and dive into our [demo](https://github.com/wwhenxuan/PySDKit/blob/main/example/demo.ipynb) to kickstart your experience—quick and easy! 🥰

pip install pysdkit

Why Develop PySDKit?

While wavelet transforms have seen remarkable integration with deep neural networks in recent years, signal decomposition techniques - hailed as the most groundbreaking time-frequency analysis method of the 21st century since the advent of Hilbert-Huang Transform (HHT) - remain significantly underutilized in deep learning applications. This gap primarily stems from the absence of a unified Python implementation framework.

To empower seamless integration of signal decomposition algorithms with deep learning architectures, accelerate research workflows, and enhance practical usability, we present PySDKit. This comprehensive library implements mainstream decomposition methodologies including:

• EMD (Empirical Mode Decomposition)
• EWT (Empirical Wavelet Transform)
• VMD (Variational Mode Decomposition)
  ...and more, serving as your essential toolkit for next-generation signal analysis.

Current Situation

PySDKit is currently developed by a single contributor (myself). While automated testing has been implemented, bugs may still exist. We warmly welcome:

1. Issue Reporting - Please open a GitHub ticket for any errors you encounter.
2. Algorithm Suggestions - Recommend promising signal decomposition methodologies via Discussions. High-value submissions will be prioritized for Python implementation.

Your contributions help shape this toolkit into a robust resource for the signal processing community.

All questions can be contacted by email: whenxuan@ieee.org

Or ask a question on Github directly: [Issues](https://github.com/wwhenxuan/PySDKit/issues)

Future Work

Currently, two of my friends will participate in the development of the project together, and we will reproduce all the algorithms listed in the table. We will also build more complete documents to help users understand the internal details and specific usage of each signal decomposition algorithm.