Necessary and sufficient conditions for causal feature selection in time series with latent common causes

Atalanti A Mastakouri, Bernhard Schölkopf, Dominik Janzing
Proceedings of the 38th International Conference on Machine Learning, PMLR 139:7502-7511, 2021.

Abstract

We study the identification of direct and indirect causes on time series with latent variables, and provide a constrained-based causal feature selection method, which we prove that is both sound and complete under some graph constraints. Our theory and estimation algorithm require only two conditional independence tests for each observed candidate time series to determine whether or not it is a cause of an observed target time series. Furthermore, our selection of the conditioning set is such that it improves signal to noise ratio. We apply our method on real data, and on a wide range of simulated experiments, which yield very low false positive and relatively low false negative rates.

Cite this Paper

BibTeX
@InProceedings{pmlr-v139-mastakouri21a, title = {Necessary and sufficient conditions for causal feature selection in time series with latent common causes}, author = {Mastakouri, Atalanti A and Sch{\"o}lkopf, Bernhard and Janzing, Dominik}, booktitle = {Proceedings of the 38th International Conference on Machine Learning}, pages = {7502--7511}, year = {2021}, editor = {Meila, Marina and Zhang, Tong}, volume = {139}, series = {Proceedings of Machine Learning Research}, month = {18--24 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v139/mastakouri21a/mastakouri21a.pdf}, url = {https://proceedings.mlr.press/v139/mastakouri21a.html}, abstract = {We study the identification of direct and indirect causes on time series with latent variables, and provide a constrained-based causal feature selection method, which we prove that is both sound and complete under some graph constraints. Our theory and estimation algorithm require only two conditional independence tests for each observed candidate time series to determine whether or not it is a cause of an observed target time series. Furthermore, our selection of the conditioning set is such that it improves signal to noise ratio. We apply our method on real data, and on a wide range of simulated experiments, which yield very low false positive and relatively low false negative rates.} }
Endnote
%0 Conference Paper %T Necessary and sufficient conditions for causal feature selection in time series with latent common causes %A Atalanti A Mastakouri %A Bernhard Schölkopf %A Dominik Janzing %B Proceedings of the 38th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Marina Meila %E Tong Zhang %F pmlr-v139-mastakouri21a %I PMLR %P 7502--7511 %U https://proceedings.mlr.press/v139/mastakouri21a.html %V 139 %X We study the identification of direct and indirect causes on time series with latent variables, and provide a constrained-based causal feature selection method, which we prove that is both sound and complete under some graph constraints. Our theory and estimation algorithm require only two conditional independence tests for each observed candidate time series to determine whether or not it is a cause of an observed target time series. Furthermore, our selection of the conditioning set is such that it improves signal to noise ratio. We apply our method on real data, and on a wide range of simulated experiments, which yield very low false positive and relatively low false negative rates.
APA
Mastakouri, A.A., Schölkopf, B. & Janzing, D.. (2021). Necessary and sufficient conditions for causal feature selection in time series with latent common causes. Proceedings of the 38th International Conference on Machine Learning, in Proceedings of Machine Learning Research 139:7502-7511 Available from https://proceedings.mlr.press/v139/mastakouri21a.html.

Related Material