Detecting causal associations in large nonlinear time series datasets
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 sam...
| Autori principali: | , , , , |
|---|---|
| Natura: | Journal article |
| Pubblicazione: |
American Association for the Advancement of Science
2019
|
| _version_ | 1826281105149394944 |
|---|---|
| author | Runge, J Nowack, P Kretschmer, M Flaxman, S Sejdinovic, D |
| author_facet | Runge, J Nowack, P Kretschmer, M Flaxman, S Sejdinovic, D |
| author_sort | Runge, J |
| collection | OXFORD |
| description | 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. |
| first_indexed | 2024-03-07T00:23:47Z |
| format | Journal article |
| id | oxford-uuid:7d6eb6ee-71b3-44c1-8d90-0f59ba723eff |
| institution | University of Oxford |
| last_indexed | 2024-03-07T00:23:47Z |
| publishDate | 2019 |
| publisher | American Association for the Advancement of Science |
| record_format | dspace |
| spelling | oxford-uuid:7d6eb6ee-71b3-44c1-8d90-0f59ba723eff2022-03-26T21:03:35ZDetecting causal associations in large nonlinear time series datasetsJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:7d6eb6ee-71b3-44c1-8d90-0f59ba723effSymplectic Elements at OxfordAmerican Association for the Advancement of Science2019Runge, JNowack, PKretschmer, MFlaxman, SSejdinovic, DIdentifying 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. |
| spellingShingle | Runge, J Nowack, P Kretschmer, M Flaxman, S Sejdinovic, D Detecting causal associations in large nonlinear time series datasets |
| title | Detecting causal associations in large nonlinear time series datasets |
| title_full | Detecting causal associations in large nonlinear time series datasets |
| title_fullStr | Detecting causal associations in large nonlinear time series datasets |
| title_full_unstemmed | Detecting causal associations in large nonlinear time series datasets |
| title_short | Detecting causal associations in large nonlinear time series datasets |
| title_sort | detecting causal associations in large nonlinear time series datasets |
| work_keys_str_mv | AT rungej detectingcausalassociationsinlargenonlineartimeseriesdatasets AT nowackp detectingcausalassociationsinlargenonlineartimeseriesdatasets AT kretschmerm detectingcausalassociationsinlargenonlineartimeseriesdatasets AT flaxmans detectingcausalassociationsinlargenonlineartimeseriesdatasets AT sejdinovicd detectingcausalassociationsinlargenonlineartimeseriesdatasets |