Joseph Lizier -- Software
Home > Software
I have developed and released the following software:
Open source project providing a stand-alone, open-source code Java implementation (usable in Matlab, Octave and Python) of information-theoretic measures of distributed computation in complex systems: i.e. information storage, transfer and modification. The distribution is available here or you can grab all the source code via github.
This toolkit includes Java implementation of all of the relevant information dynamics measures, including basic measures such as entropy, joint entropy, mutual information, conditional mutual information, as well as advanced measures such as transfer entropy, conditional transfer entropy, active information storage, excess entropy, separable information.
All measures are implemented using several different estimation techniques, including discrete-valued estimators, box-kernel estimation, Kraskov-Grassberger estimators, symbolic coding, and Gaussian approximation.
This standalone toolkit can be used as a Java plugin in Octave, Matlab and Python (see details on google code site).
Please note that this toolkit replaces the previous Matlab transfer entropy code that was referred to in some of my earlier presentations.
A a comprehensive python package for efficient inference of networks and their node dynamics from multivariate time series data using information theory, in particular using multivariate transfer entropy. Read more and download from github.
A simple framework for running processes on a cluster using qsub commands. Read more and download from here.
Related software from others:
- TRENTOOL - the Transfer Entropy Toolbox for the Fieldtrip data format, from Michael Lindner, Raul Vicente and Michael Wibral. TRENTOOL provides an implementation of the transfer entropy (harnessing the Kraskov variable kernel width algorithm to handle small data sets) which is guided towards effective network inference in neural data sets.