tefs

Causal feature selection for time series data using transfer entropy


Keywords
tefs, transfer, entropy, feature, selection, causality, time, series
License
Other
Install
pip install tefs==0.3.0

Documentation

Transfer Entropy Feature Selection

License: MIT PyPI - Version PyPI - Downloads

This repository implements a causal feature selection algorithm based on transfer entropy. The algorithm is described in: Bonetti, P., Metelli, A. M., & Restelli, M. (2023, October 17). Causal Feature Selection via Transfer Entropy. (https://arxiv.org/abs/2310.11059).

Installation

The package can be installed using pip:

pip install tefs

How to use

Refer to the documentation for usage examples.

Attribution

If you use this package in your research, please cite the following paper:

@misc{bonetti2023causal,
    title = {Causal {{Feature Selection}} via {{Transfer Entropy}}},
    author = {Bonetti, Paolo and Metelli, Alberto Maria and Restelli, Marcello},
    year = 2023,
    eprint = {2310.11059},
    archiveprefix = {arXiv},
    primaryclass = {cs.LG}
}

Authors