Assessing Nonlinear Granger Causality from Multivariate Time Series
PDF PDF
A straightforward nonlinear extension of Grangers concept of causality in the kernel framework is suggested. The kernel-based approach to assessing nonlinear Granger causality in multivariate time series enables us to determine, in a model-free way, whether the causal relation between two time series is present or not and whether it is direct or mediated by other processes. The trace norm of the so-called covariance operator in feature space is used to measure the prediction error. Relying on this measure, we test the improvement of predictability between time series by subsampling-based multiple testing. The distributional properties of the resulting p-values reveal the direction of Granger causality. Experiments with simulated and real-world data show that our method provides encouraging results.
| Author(s): | Sun, X. |
| Links: | |
| Journal: | Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2008 |
| Pages: | 440-455 |
| Year: | 2008 |
| Month: | September |
| Day: | 0 |
| Editors: | Daelemans, W. , B. Goethals, K. Morik |
| Publisher: | Springer |
| BibTeX Type: | Conference Paper (inproceedings) |
| Address: | Berlin, Germany |
| DOI: | 10.1007/978-3-540-87481-2_29 |
| Event Name: | European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2008) |
| Event Place: | Antwerpen, Belgium |
| Digital: | 0 |
| Electronic Archiving: | grant_archive |
| Language: | en |
| Organization: | Max-Planck-Gesellschaft |
| School: | Biologische Kybernetik |
BibTeX
@inproceedings{5254,
title = {Assessing Nonlinear Granger Causality from Multivariate Time Series},
journal = {Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2008},
abstract = {A straightforward nonlinear extension of Grangers concept of causality in the kernel framework is suggested. The kernel-based approach to assessing nonlinear Granger causality in multivariate time series enables us to determine, in a model-free way, whether the causal relation between two time series is present or not and whether it is direct or mediated by other processes. The trace norm of the so-called covariance operator in feature space is used to measure the prediction error. Relying on this measure, we test the improvement of predictability between time series by subsampling-based multiple testing. The distributional properties of the resulting p-values reveal the direction of Granger causality. Experiments with simulated and real-world data show that our method provides encouraging results.},
pages = {440-455},
editors = {Daelemans, W. , B. Goethals, K. Morik},
publisher = {Springer},
organization = {Max-Planck-Gesellschaft},
school = {Biologische Kybernetik},
address = {Berlin, Germany},
month = sep,
year = {2008},
author = {Sun, X.},
doi = {10.1007/978-3-540-87481-2_29},
month_numeric = {9}
}