Skip to content

Latest commit

 

History

History
73 lines (49 loc) · 2.66 KB

README.rst

File metadata and controls

73 lines (49 loc) · 2.66 KB
Raw

The iisignature package

This package provides tools for calculating the signature and log signature of a data stream. These are summary statistics of piecewise linear paths, consisting of iterated integrals, inspired by rough path theory. See <https://github.com/bottler/phd-docs/blob/master/iisignature.pdf> for documentation, the page <https://github.com/bottler/phd-docs> for more information about the calculations, and <https://github.com/bottler/iisignature> for source code.

It is work in progress.

Python

Install with:

pip install iisignature

Don't install this package if you don't have numpy > 1.7 installed. On Windows, this package is not usable with Python 2.x. For Python 3.5, 3.6, 3.7 and 3.8 on Windows, there are precompiled wheels available, you may need to do pip install wheel to use them. On other platforms, you will need to be able to compile C++ extensions. On a Mac, you will need to have installed Xcode and the Xcode command line tools before doing the installation. The fastest, on-the-fly compiled versions of the log signature calculations are for Windows, Mac and Linux, on both 32 and 64 bit.

Sometimes there are compile errors on a Mac, if this happens try:

MACOSX_DEPLOYMENT_TARGET=10.9 pip install iisignature

Use without installing a released version

To install the current "master" development version straight from github you can type:

pip install git+git://github.com/bottler/iisignature

From a checkout of this repository, you can build the extension and run the tests using:

python setup.py test

From a checkout of this repository, you can build the extension into this directory itself using:

python setup.py build_ext --inplace

after which you can use the package in Python in this directory, use the examples in the examples directory, or add this directory to your PYTHONPATH and then use the package in Python anywhere.

Paper

A paper in TOMS <https://dl.acm.org/doi/10.1145/3371237> accompanies the library. It can also be found on arxiv at <https://arxiv.org/abs/1802.08252>. If you find iisignature useful in your research then please cite:

@article{iisignature,
title={Algorithm 1004: The iisignature Library:

Efficient Calculation of Iterated-Integral Signatures and Log Signatures},

author={Jeremy Reizenstein and Benjamin Graham}, journal={ACM Transactions on Mathematical Software (TOMS)}, year={2020}

}

Authors

  • Benjamin Graham
  • Jeremy Reizenstein

Thanks

This package includes BCH coefficients from Fernando Casas and Ander Murua. <http://arxiv.org/abs/0810.2656>