sax-ts 1.0.0

Symbolic Aggregate approXimation (SAX)

Symbolic Aggregate approXimation

This package implements the Symbolic Aggregate approXimation (SAX) algorithm in C++ with Python bindings using pybind11. SAX is a time series discretization method that represents a continuous-valued time series as a (sliding window of) symbolic representation(s).

Jessica Lin's SAX page

Eamonn Keogh's SAX page

Installation

Currently, only building from source is available. To build from source, make sure you have a C++ compiler and the header files for your Python version installed. Download/clone the directory, change into the directory, then install from source using pip install ., or, for testing pip install ".[test]".

The project is set up so that pybind11 is installed as a build dependency. The core build has no other C++ or Python dependencies.

You can add this to your own project's requirements with the following entry in your pyproject.toml file:

dependencies = [
    "sax-ts @ git+https://github.com/steenrotsman/sax-ts.git",
]

References

P. Patel, E. Keogh, J. Lin and S. Lonardi, "Mining motifs in massive time series databases," 2002 IEEE International Conference on Data Mining, 2002. Proceedings., Maebashi City, Japan, 2002, pp. 370-377, doi: 10.1109/ICDM.2002.1183925.

J. Lin, E. Keogh, S. Lonardi, and B. Chiu, "A symbolic representation of time series, with implications for streaming algorithms," In Proceedings of the 8th ACM SIGMOD workshop on Research issues in data mining and knowledge discovery, 2003, pp. 2-11, doi: 10.1145/882082.882086.

J. Lin, E. Keogh, L. Wei, and S. Lonardi, "Experiencing SAX: a novel symbolic representation of time series," Data Min Knowl Disc, vol. 15, pp. 107–144, Apr. 2007, doi: 10.1007/s10618-007-0064-z.