- Keywords: time series, causality, missing data, chaos
- TL;DR: Methodology to infer causal dependencies between irregularly sampled time series with missing values.
- Abstract: Discovering causal structures of processes is a major tool of scientific inquiry because it helps better understand and explain the mechanisms driving a phenomenon of interest. However, accurately inferring causal structures based on observational data only is still an open problem. In particular, this problem becomes increasingly difficult when it relies on data with missing values. In this article, we present a method to uncover causal relations between chaotic dynamical systems from sporadic time series (that is, incomplete observations at infrequent and irregular intervals), which builds upon Convergent Cross Mapping and recent advances in continuous time-series modeling (GRU-ODE-Bayes).