Description

A particular focus in the XAIDA-related developments was the implementation of mixed-type non-parametric conditional independence tests applicable also to synergistic problems, non-stationary causal discovery methods, and non-parametric causal mediation methods for the estimation of direct and indirect effects in mixed-type data; mediation analysis can also inform about which pathways a particular effect should be attributed to.

Potential Users

• Researchers, in applied packages of XAIDA or otherwise concerned with causal relations in climate (or similar) data.
• Technically interested planners or engineers who wish to quantify the effect of interventions or supplement their causal understanding of a problem.

Guide: An extensive guide and many tutorials can be found in the GitHub repository https://github.com/jakobrunge/tigramite/.

Availability: The tool is a pip-installable Python package. It can be used with basic knowledge of the Python programming language.

Use Cases: For a list of in-depth tutorials as well as use cases see https://github.com/jakobrunge/tigramite/tree/master/tutorials/.

References

• Detecting causal associations in large nonlinear time series datasets. J. Runge, P. Nowack, M. Kretschmer, S. Flaxman, and D. Sejdinovic. Science Advances 5(11): eaau4996 (2019).
• Causal network reconstruction from time series: From theoretical assumptions to practical estimation. J. Runge. Chaos: An Interdisciplinary Journal of Nonlinear Science, 28:7 (2018).