Detecting causal associations in large nonlinear time series datasets

2017/02/28 by Jakob Runge, Peer Nowack, Marlene Kretschmer +2 · 3 citations
Mathematics · Physics and Astronomy · #stat.ME #physics.ao-ph #stat.AP

paper · pdf · doi:10.1126/sciadv.aau4996

published as Science Advances Vol. 5, no. 11, eaau4996 (2019) · 46 pages, 19 figures

arxiv created 2018/06/28 · arxiv updated 2019/12/03

Abstract

Identifying causal relationships from observational time series data is a key problem in disciplines such as climate science or neuroscience, where experiments are often not possible. Data-driven causal inference is challenging since datasets are often high-dimensional and nonlinear with limited sample sizes. Here we introduce a novel method that flexibly combines linear or nonlinear conditional independence tests with a causal discovery algorithm that allows to reconstruct causal networks from large-scale time series datasets. We validate the method on a well-established climatic teleconnection connecting the tropical Pacific with extra-tropical temperatures and using large-scale synthetic datasets mimicking the typical properties of real data. The experiments demonstrate that our method outperforms alternative techniques in detection power from small to large-scale datasets and opens up entirely new possibilities to discover causal networks from time series across a range of research fields.

Cited by

• Style-aware gloss control for generative non-photorealistic rendering
• Semiparametric Causal Discovery and Inference with Invalid Instruments
• A conceptual synthesis of causal assumptions for causal discovery and inference
• Emergence of psychopathological computations in large language models