Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/blue-yonder/tsfresh
Automatic extraction of relevant features from time series:
https://github.com/blue-yonder/tsfresh
data-science feature-extraction time-series
Last synced: 3 days ago
JSON representation
Automatic extraction of relevant features from time series:
- Host: GitHub
- URL: https://github.com/blue-yonder/tsfresh
- Owner: blue-yonder
- License: mit
- Created: 2016-10-26T11:29:17.000Z (about 8 years ago)
- Default Branch: main
- Last Pushed: 2024-04-22T20:07:04.000Z (8 months ago)
- Last Synced: 2024-05-15T13:56:12.861Z (7 months ago)
- Topics: data-science, feature-extraction, time-series
- Language: Jupyter Notebook
- Homepage: http://tsfresh.readthedocs.io
- Size: 7.97 MB
- Stars: 8,125
- Watchers: 168
- Forks: 1,200
- Open Issues: 67
-
Metadata Files:
- Readme: README.md
- Changelog: CHANGES.rst
- License: LICENSE.txt
- Authors: AUTHORS.rst
Awesome Lists containing this project
- awesome-AIOps - 时间序列特征提取库tsfresh
- awesome-systematic-trading - tsfresh - yonder/tsfresh) |  | (TimeSeries Analysis / Cryptocurrencies)
- awesome_time_series_in_python - tsfresh
- awesome-systematic-trading - tsfresh - Automatic extraction of relevant features from time series. (Analytic tools / TimeSeries Analysis)
- awesome-time-series - tsfresh
- my-awesome-github-stars - blue-yonder/tsfresh - Automatic extraction of relevant features from time series: (Jupyter Notebook)
- awesome_quantmetry - tsfresh
- awesome-python-machine-learning - tsfresh - Automatic extraction of relevant features from time series. (Uncategorized / Uncategorized)
- Awesome-time-series - 📚 tsfresh - series feature extraction by combining established algorithms from statistics, time-series analysis, signal processing, and nonlinear dynamics with a robust feature selection algorithm. In this context, the term time-series is interpreted in the broadest possible sense, such that any types of sampled data or even event sequences can be characterised. (Others)
- awesome-sciml - blue-yonder/tsfresh: Automatic extraction of relevant features from time series:
- awesome-quant - tsfresh - Automatic extraction of relevant features from time series. (Python / Time Series)
- awesome-failure-diagnosis - tsfresh
- awesome-marketing-machine-learning - tsfresh - yonder/tsfresh.svg?style=social) (Time Series)
- awesome-list - TSFRESH - Automatic extraction of relevant features from time series. (Data Processing / Data Pre-processing & Loading)
- awesome-quant - tsfresh - Automatic extraction of relevant features from time series. (Python / Time Series)
- awesome-production-machine-learning - tsfresh - yonder/tsfresh.svg?style=social) - Automatic extraction of relevant features from time series. (AutoML)
- awesome-mlops - TSFresh - Python library for automatic extraction of relevant features from time series. (Feature Engineering)
- awesome-python-machine-learning-resources - GitHub - 10% open · ⏱️ 21.12.2021): (时间序列)
- StarryDivineSky - blue-yonder/tsfresh
- pytrade.org - tsfresh - Automatic extraction of relevant features from time series (Curated List / Time Series Analysis)
README
-----------------
# tsfresh
[](https://tsfresh.readthedocs.io/en/latest/?badge=latest)
[](https://github.com/blue-yonder/tsfresh/actions)
[](https://codecov.io/gh/blue-yonder/tsfresh)
[](https://github.com/blue-yonder/tsfresh/blob/main/LICENSE.txt)
[](https://mybinder.org/v2/gh/blue-yonder/tsfresh/main?filepath=notebooks)
[](https://pepy.tech/project/tsfresh)
This repository contains the *TSFRESH* python package. The abbreviation stands for
*"Time Series Feature extraction based on scalable hypothesis tests"*.
The package provides systematic time-series feature extraction by combining established algorithms from statistics, time-series analysis, signal processing, and nonlinear dynamics with a robust feature selection algorithm. In this context, the term *time-series* is interpreted in the broadest possible sense, such that any types of sampled data or even event sequences can be characterised.
## Spend less time on feature engineering
Data Scientists often spend most of their time either cleaning data or building features.
While we cannot change the first thing, the second can be automated.
*TSFRESH* frees your time spent on building features by extracting them automatically.
Hence, you have more time to study the newest deep learning paper, read hacker news or build better models.
## Automatic extraction of 100s of features
*TSFRESH* automatically extracts 100s of features from time series.
Those features describe basic characteristics of the time series such as the number of peaks, the average or maximal value or more complex features such as the time reversal symmetry statistic.
The set of features can then be used to construct statistical or machine learning models on the time series to be used for example in regression or
classification tasks.
## Forget irrelevant features
Time series often contain noise, redundancies or irrelevant information.
As a result most of the extracted features will not be useful for the machine learning task at hand.
To avoid extracting irrelevant features, the *TSFRESH* package has a built-in filtering procedure.
This filtering procedure evaluates the explaining power and importance of each characteristic for the regression or classification tasks at hand.
It is based on the well developed theory of hypothesis testing and uses a multiple test procedure.
As a result the filtering process mathematically controls the percentage of irrelevant extracted features.
The *TSFRESH* package is described in the following open access paper:
* Christ, M., Braun, N., Neuffer, J., and Kempa-Liehr A.W. (2018).
_Time Series FeatuRe Extraction on basis of Scalable Hypothesis tests (tsfresh -- A Python package)._
Neurocomputing 307, p. 72-77, [doi: 10.1016/j.neucom.2018.03.067](https://doi.org/10.1016/j.neucom.2018.03.067).
The FRESH algorithm is described in the following whitepaper:
* Christ, M., Kempa-Liehr, A.W., and Feindt, M. (2017).
_Distributed and parallel time series feature extraction for industrial big data applications._
ArXiv e-print 1610.07717, [https://arxiv.org/abs/1610.07717](https://arxiv.org/abs/1610.07717).
Systematic time-series feature extraction even works for unsupervised problems:
* Teh, H.Y., Wang, K.I-K., Kempa-Liehr, A.W. (2021).
_Expect the Unexpected: Unsupervised feature selection for automated sensor anomaly detection._
IEEE Sensors Journal 15.16, p. 18033-18046, [doi: 10.1109/JSEN.2021.3084970](https://doi.org/10.1109/JSEN.2021.3084970).
Due to the fact that tsfresh basically provides time-series feature extraction for free, you can now concentrate on engineering new time-series,
like e.g. differences of signals from synchronous measurements, which provide even better time-series features:
* Kempa-Liehr, A.W., Oram, J., Wong, A., Finch, M., Besier, T. (2020).
_Feature engineering workflow for activity recognition from synchronized inertial measurement units._
In: Pattern Recognition. ACPR 2019. Ed. by M. Cree et al. Vol. 1180. Communications in Computer and Information Science (CCIS).
Singapore: Springer, p. 223–231. [doi: 10.1007/978-981-15-3651-9_20](https://doi.org/10.1007/978-981-15-3651-9_20).
* Simmons, S., Jarvis, L., Dempsey, D., Kempa-Liehr, A.W. (2021).
_Data Mining on Extremely Long Time-Series._
In: 2021 International Conference on Data Mining Workshops (ICDMW). Ed. by B. Xue et al.
Los Alamitos: IEEE, p. 1057-1066. [doi: 10.1109/ICDMW53433.2021.00137](https://doi.org/10.1109/ICDMW53433.2021.00137).
Systematic time-series features engineering allows to work with time-series samples of different lengths, because every time-series is projected
into a well-defined feature space. This approach allows the design of robust machine learning algorithms in applications with missing data.
* Kennedy, A., Gemma, N., Rattenbury, N., Kempa-Liehr, A.W. (2021).
_Modelling the projected separation of microlensing events using systematic time-series feature engineering._
Astronomy and Computing 35.100460, p. 1–14, [doi: 10.1016/j.ascom.2021.100460](https://doi.org/10.1016/j.ascom.2021.100460)
Is your time-series classification problem imbalanced? There is a good chance that undersampling of time-series feature matrices
might solve your problem:
* Dempsey, D.E., Cronin, S.J., Mei, S., Kempa-Liehr, A.W. (2020).
_Automatic precursor recognition and real-time forecasting of sudden explosive volcanic eruptions at Whakaari, New Zealand_.
Nature Communications 11.3562, p. 1-8, [doi: 10.1038/s41467-020-17375-2](https://doi.org/10.1038/s41467-020-17375-2).
Natural language processing of written texts is an example of applying systematic time-series feature engineering to event sequences,
which is described in the following open access paper:
* Tang, Y., Blincoe, K., Kempa-Liehr, A.W. (2020).
_Enriching Feature Engineering for Short Text Samples by Language Time Series Analysis._
EPJ Data Science 9.26, p. 1–59. [doi: 10.1140/epjds/s13688-020-00244-9](https://doi.org/10.1140/epjds/s13688-020-00244-9)
## Advantages of tsfresh
*TSFRESH* has several selling points, for example
1. it is field tested
2. it is unit tested
3. the filtering process is statistically/mathematically correct
4. it has a comprehensive documentation
5. it is compatible with sklearn, pandas and numpy
6. it allows anyone to easily add their favorite features
7. it both runs on your local machine or even on a cluster
## Next steps
If you are interested in the technical workings, go to see our comprehensive Read-The-Docs documentation at [http://tsfresh.readthedocs.io](http://tsfresh.readthedocs.io).
The algorithm, especially the filtering part are also described in the paper mentioned above.
We appreciate any contributions, if you are interested in helping us to make *TSFRESH* the biggest archive of feature extraction methods in python, just head over to our [How-To-Contribute](http://tsfresh.readthedocs.io/en/latest/text/how_to_contribute.html) instructions.
If you want to try out `tsfresh` quickly or if you want to integrate it into your workflow, we also have a docker image available:
docker pull nbraun/tsfresh
## Backwards compatibility
If you need to reproduce or update time-series features, which were computed with the `matrixprofile` feature calculators, you need to create a Python 3.8 environment:
conda create --name tsfresh__py_3.8 python=3.8
conda activate tsfresh__py_3.8
pip install tsfresh[matrixprofile]
## Acknowledgements
The research and development of *TSFRESH* was funded in part by the German Federal Ministry of Education and Research under grant number 01IS14004 (project iPRODICT).