v0.0.11

What's Changed

• whenxuan: fix the nan error in forecast_pfn by @wwhenxuan in #45

Full Changelog: v0.0.10...v0.0.11

v0.0.10

What's Changed

• whenxuan: update the time series augmentation and unittest by @wwhenxuan in #43
• whenxuan: update the verion for 0.0.10 by @wwhenxuan in #44

Full Changelog: v0.0.9...v0.0.10

v0.0.9

What's Changed

• whenxuan: update the time series augmentation by @wwhenxuan in #39
• whenxuan: update the version for 0.0.9 by @wwhenxuan in #42

Full Changelog: v0.0.8...v0.0.9

v0.0.8

What's Changed

• whenxuan: update the v0.0.8 version for s2generator by @wwhenxuan in #38

Full Changelog: https://github.com/wwhenxuan/S2Generator/compare/v0.0.7...v0## What's Changed

• whenxuan: update the v0.0.8 version for s2generator by @wwhenxuan in #38

Full Changelog: v0.0.7...v0.0.8.0.8

v0.0.7

Based on the AR parameterized spectrum estimation method in modern signal processing, we construct a novel learnable time series generator using Wiener filters. Compared to the previous generation of ARIMA models, the Wiener filter can fit stationary time series at a faster speed.

What's Changed

• whenxuan: update the wiener filter for simulate by @wwhenxuan in #37

Full Changelog: v0.0.6...v0.0.7

v0.0.6

What's Changed

• Fixed a bug in the ARIMA model caused by linear operations. by @wwhenxuan in #35
• whenxuan: update the v0.0.6 version for s2generator by @wwhenxuan in #36

Full Changelog: v0.0.5...v0.0.6

v0.0.5

Considering the significant randomness inherent in generating response sequences for complex systems simply by constructing symbolic expressions and excitation time series, we further develop a learnable data generation method. We believe that all stationary signals can be obtained by exciting a linear time-invariant system with white noise. Furthermore, the difference operation can transform non-stationary signals into stationary ones. Therefore, we construct a linear system with a difference form using an ARIMA model. This system can learn the statistical representation of the input sequence, especially its autocorrelation and power spectral density, thereby generating time series.

What's Changed

• Add the code and example for the arima simulator by @wwhenxuan in #33
• whenxuan: update the version by @wwhenxuan in #34

Full Changelog: v0.0.4...v0.0.5

v0.0.4

S2Generator 0.0.4

We're happy to announce the release of S2Generator 0.0.4! 🎉🎉🎉

What is the S2Generator

Foundation models for time series analysis (TSA) have attracted significant attention. However, challenges such as training data scarcity and imbalance continue to hinder their development. Inspired by complex dynamic system theories, we design a series-symbol ($S^2$) data generation mechanism, enabling the unrestricted creation of high-quality time series data paired with corresponding symbolic expressions 🥳.

The Usage and Installation of S2Generator

This is the first version of our $S^2$ data generation mechanism that is open source and officially released. We developed it using only data science libraries such as NumPy, Scipy, and Matplotlib. You can install it via pip:

pip install s2generator

For more specific usage, please see our Demo file ✨.

If you encounter any issues while using S2Generator, please contact us immediately. We would be very grateful.

Future Work for S2Generator

We will also accelerate the development of S2Generator's technical documentation, providing more comprehensive and diverse examples and usage instructions 📃.

If you find this $S^2$ data generation method helpful, please cite the following paper:

@misc{wang2025mitigatingdatascarcitytime,
      title={Mitigating Data Scarcity in Time Series Analysis: A Foundation Model with Series-Symbol Data Generation}, 
      author={Wenxuan Wang and Kai Wu and Yujian Betterest Li and Dan Wang and Xiaoyu Zhang and Jing Liu},
      year={2025},
      eprint={2502.15466},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2502.15466}, 
}

Full Changelog: v0.0.2...v0.0.3

What's Changed

• Update the docs for images and examples by @wwhenxuan in #26
• whenxuan: adjust the images by @wwhenxuan in #27
• whenxuan: adjust the images by @wwhenxuan in #28
• rename by @wwhenxuan in #29
• whenxuan: rename the pip by @wwhenxuan in #30
• whenxuan: fix the docs and unittest by @wwhenxuan in #32

Full Changelog: v0.0.3...v0.0.4

v0.0.3

S2Generator 0.0.3

We're happy to announce the release of S2Generator 0.0.3! 🎉🎉🎉

What is the S2Generator

Foundation models for time series analysis (TSA) have attracted significant attention. However, challenges such as training data scarcity and imbalance continue to hinder their development. Inspired by complex dynamic system theories, we design a series-symbol ($S^2$) data generation mechanism, enabling the unrestricted creation of high-quality time series data paired with corresponding symbolic expressions 🥳.

The Usage and Installation of S2Generator

This is the first version of our $S^2$ data generation mechanism that is open source and officially released. We developed it using only data science libraries such as NumPy, Scipy, and Matplotlib. You can install it via pip:

pip install s2generator

For more specific usage, please see our Demo file ✨.

If you encounter any issues while using S2Generator, please contact us immediately. We would be very grateful.

Future Work for S2Generator

We will also accelerate the development of S2Generator's technical documentation, providing more comprehensive and diverse examples and usage instructions 📃.

If you find this $S^2$ data generation method helpful, please cite the following paper:

@misc{wang2025mitigatingdatascarcitytime,
      title={Mitigating Data Scarcity in Time Series Analysis: A Foundation Model with Series-Symbol Data Generation}, 
      author={Wenxuan Wang and Kai Wu and Yujian Betterest Li and Dan Wang and Xiaoyu Zhang and Jing Liu},
      year={2025},
      eprint={2502.15466},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2502.15466}, 
}

Full Changelog: v0.0.2...v0.0.3

v0.0.2

S2Generator 0.0.2

We're happy to announce the release of S2Generator 0.0.2! 🎉🎉🎉

What is the S2Generator

Foundation models for time series analysis (TSA) have attracted significant attention. However, challenges such as training data scarcity and imbalance continue to hinder their development. Inspired by complex dynamic system theories, we design a series-symbol ($S^2$) data generation mechanism, enabling the unrestricted creation of high-quality time series data paired with corresponding symbolic expressions 🥳.

The Usage and Installation of S2Generator

This is the first version of our $S^2$ data generation mechanism that is open source and officially released. We developed it using only data science libraries such as NumPy, Scipy, and Matplotlib. You can install it via pip:

pip install s2generator

For more specific usage, please see our Demo file ✨.

If you encounter any issues while using S2Generator, please contact us immediately. We would be very grateful.

Future Work for S2Generator

We will also accelerate the development of S2Generator's technical documentation, providing more comprehensive and diverse examples and usage instructions 📃.

If you find this $S^2$ data generation method helpful, please cite the following paper:

@misc{wang2025mitigatingdatascarcitytime,
      title={Mitigating Data Scarcity in Time Series Analysis: A Foundation Model with Series-Symbol Data Generation}, 
      author={Wenxuan Wang and Kai Wu and Yujian Betterest Li and Dan Wang and Xiaoyu Zhang and Jing Liu},
      year={2025},
      eprint={2502.15466},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2502.15466}, 
}

What's Changed

• complete diff operator by @johnfan12 in #2
• Master by @wwhenxuan in #3
• Merge pull request #3 from wwhenxuan/master by @wwhenxuan in #4
• Add ARMA-stimulated time series generation module by @wwhenxuan in #5
• Update new time series generator by @wwhenxuan in #6
• Master by @wwhenxuan in #7
• Completed All Incentive Generation by @wwhenxuan in #8
• Added a new parameter control module by @wwhenxuan in #9
• Completed the first major update by @wwhenxuan in #10
• Update the wasserstein distance by @wwhenxuan in #11
• Add the Print Status for S2Generator by @wwhenxuan in #12
• Add the Save and Load Function for S2 data by @wwhenxuan in #13
• Update the MAIN for S2Generation by @wwhenxuan in #14
• Add the examples and utils function by @wwhenxuan in #15
• Update More Decomposition Methods for Time Series Analysis by @wwhenxuan in #16
• Test excite by @johnfan12 in #17
• Add the unit test for the Exctiation by @wwhenxuan in #18
• Update the Logo by @wwhenxuan in #19
• Update the README by @wwhenxuan in #20
• whenxuan: update pypi version by @wwhenxuan in #21
• Document construction by @changewam in #22
• whenxuan: fix the ignore for docs by @wwhenxuan in #23
• Change thumbnail in gallery by @changewam in #24

New Contributors

• @wwhenxuan made their first contribution in #3
• @changewam made their first contribution in #22

Full Changelog: https://github.com/wwhenxuan/S2Generator/commits/v0.0.2