Necessary and sufficient conditions for causal feature selection in time series with latent common causes
This addresses causal feature selection in time series for researchers and practitioners dealing with latent variables, offering a more reliable alternative to existing methods like Granger causality.
The paper tackles the problem of identifying direct and indirect causes in time series with latent common causes, providing necessary and sufficient conditions under graph constraints. The method requires only two conditional independence tests per candidate time series and outperforms Granger causality with very low false positives and relatively low false negatives in simulations and real data.
We study the identification of direct and indirect causes on time series and provide conditions in the presence of latent variables, which we prove to be necessary and sufficient under some graph constraints. Our theoretical results and estimation algorithms require 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. We provide experimental results in simulations, as well as real data. Our results show that our method leads to very low false positives and relatively low false negative rates, outperforming the widely used Granger causality.