Reliability of the Granger causality inference

Douglas Zhou, Yaoyu Zhang, Yanyang Xiao, David Cai

Research output: Contribution to journalArticle

Abstract

How to characterize information flows in physical, biological, and social systems remains a major theoretical challenge. Granger causality (GC) analysis has been widely used to investigate information flow through causal interactions. We address one of the central questions in GC analysis, that is, the reliability of the GC evaluation and its implications for the causal structures extracted by this analysis. Our work reveals that the manner in which a continuous dynamical process is projected or coarse-grained to a discrete process has a profound impact on the reliability of the GC inference, and different sampling may potentially yield completely opposite inferences. This inference hazard is present for both linear and nonlinear processes. We emphasize that there is a hazard of reaching incorrect conclusions about network topologies, even including statistical (such as small-world or scale-free) properties of the networks, when GC analysis is blindly applied to infer the network topology. We demonstrate this using a small-world network for which a drastic loss of small-world attributes occurs in the reconstructed network using the standard GC approach. We further show how to resolve the paradox that the GC analysis seemingly becomes less reliable when more information is incorporated using finer and finer sampling. Finally, we present strategies to overcome these inference artifacts in order to obtain a reliable GC result.

Original languageEnglish (US)
Article number043016
JournalNew Journal of Physics
Volume16
DOIs
StatePublished - 2014

Fingerprint

inference
information flow
hazards
topology
sampling
paradoxes
artifacts
evaluation
interactions

Keywords

  • Causal information flow
  • Reliability of causal inference
  • Sampling hazards
  • Small-world topology

ASJC Scopus subject areas

  • Physics and Astronomy(all)

Cite this

Reliability of the Granger causality inference. / Zhou, Douglas; Zhang, Yaoyu; Xiao, Yanyang; Cai, David.

In: New Journal of Physics, Vol. 16, 043016, 2014.

Research output: Contribution to journalArticle

Zhou, Douglas ; Zhang, Yaoyu ; Xiao, Yanyang ; Cai, David. / Reliability of the Granger causality inference. In: New Journal of Physics. 2014 ; Vol. 16.
@article{5c668c5712ba4394ab8188fb6f287988,
title = "Reliability of the Granger causality inference",
abstract = "How to characterize information flows in physical, biological, and social systems remains a major theoretical challenge. Granger causality (GC) analysis has been widely used to investigate information flow through causal interactions. We address one of the central questions in GC analysis, that is, the reliability of the GC evaluation and its implications for the causal structures extracted by this analysis. Our work reveals that the manner in which a continuous dynamical process is projected or coarse-grained to a discrete process has a profound impact on the reliability of the GC inference, and different sampling may potentially yield completely opposite inferences. This inference hazard is present for both linear and nonlinear processes. We emphasize that there is a hazard of reaching incorrect conclusions about network topologies, even including statistical (such as small-world or scale-free) properties of the networks, when GC analysis is blindly applied to infer the network topology. We demonstrate this using a small-world network for which a drastic loss of small-world attributes occurs in the reconstructed network using the standard GC approach. We further show how to resolve the paradox that the GC analysis seemingly becomes less reliable when more information is incorporated using finer and finer sampling. Finally, we present strategies to overcome these inference artifacts in order to obtain a reliable GC result.",
keywords = "Causal information flow, Reliability of causal inference, Sampling hazards, Small-world topology",
author = "Douglas Zhou and Yaoyu Zhang and Yanyang Xiao and David Cai",
year = "2014",
doi = "10.1088/1367-2630/16/4/043016",
language = "English (US)",
volume = "16",
journal = "New Journal of Physics",
issn = "1367-2630",
publisher = "IOP Publishing Ltd.",

}

TY - JOUR

T1 - Reliability of the Granger causality inference

AU - Zhou, Douglas

AU - Zhang, Yaoyu

AU - Xiao, Yanyang

AU - Cai, David

PY - 2014

Y1 - 2014

N2 - How to characterize information flows in physical, biological, and social systems remains a major theoretical challenge. Granger causality (GC) analysis has been widely used to investigate information flow through causal interactions. We address one of the central questions in GC analysis, that is, the reliability of the GC evaluation and its implications for the causal structures extracted by this analysis. Our work reveals that the manner in which a continuous dynamical process is projected or coarse-grained to a discrete process has a profound impact on the reliability of the GC inference, and different sampling may potentially yield completely opposite inferences. This inference hazard is present for both linear and nonlinear processes. We emphasize that there is a hazard of reaching incorrect conclusions about network topologies, even including statistical (such as small-world or scale-free) properties of the networks, when GC analysis is blindly applied to infer the network topology. We demonstrate this using a small-world network for which a drastic loss of small-world attributes occurs in the reconstructed network using the standard GC approach. We further show how to resolve the paradox that the GC analysis seemingly becomes less reliable when more information is incorporated using finer and finer sampling. Finally, we present strategies to overcome these inference artifacts in order to obtain a reliable GC result.

AB - How to characterize information flows in physical, biological, and social systems remains a major theoretical challenge. Granger causality (GC) analysis has been widely used to investigate information flow through causal interactions. We address one of the central questions in GC analysis, that is, the reliability of the GC evaluation and its implications for the causal structures extracted by this analysis. Our work reveals that the manner in which a continuous dynamical process is projected or coarse-grained to a discrete process has a profound impact on the reliability of the GC inference, and different sampling may potentially yield completely opposite inferences. This inference hazard is present for both linear and nonlinear processes. We emphasize that there is a hazard of reaching incorrect conclusions about network topologies, even including statistical (such as small-world or scale-free) properties of the networks, when GC analysis is blindly applied to infer the network topology. We demonstrate this using a small-world network for which a drastic loss of small-world attributes occurs in the reconstructed network using the standard GC approach. We further show how to resolve the paradox that the GC analysis seemingly becomes less reliable when more information is incorporated using finer and finer sampling. Finally, we present strategies to overcome these inference artifacts in order to obtain a reliable GC result.

KW - Causal information flow

KW - Reliability of causal inference

KW - Sampling hazards

KW - Small-world topology

UR - http://www.scopus.com/inward/record.url?scp=84899514361&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84899514361&partnerID=8YFLogxK

U2 - 10.1088/1367-2630/16/4/043016

DO - 10.1088/1367-2630/16/4/043016

M3 - Article

VL - 16

JO - New Journal of Physics

JF - New Journal of Physics

SN - 1367-2630

M1 - 043016

ER -