Java Information Dynamics Toolkit (JIDT)
- Java Information Dynamics Toolkit (JIDT) provides a standalone, open-source code Java implementation (which is usable in Matlab, Octave and Python) of various information-theoretic measures. Primarily these are aimed at measuring how information is stored, transferred and modified in multivariate systems (e.g. via transfer entropy), but also include more standard measures such as entropy and mutual information. The measures are implemented for both discrete and continuous data, using various estimator types (e.g. kernel and Kraskov-Grassberger estimators). Demonstration code is included to show, for example, how to use the toolkit in Matlab and Python, and how to filter space-time dynamics of information in cellular automata.
- Joseph Lizier
- Information Theory, Cellular Automata, Entropy, Algorithmic Complexity