Semiparametric detection of nonlinear causal coupling using partial directed coherence.

Area of Science:

• Complex Systems Analysis
• Time Series Analysis
• Causal Inference

Background:

• Inferring causal relationships from time series data is challenging, especially with nonlinear couplings.
• Existing methods often require extensive parameterization and complex optimization, hindering practical application.

Purpose of the Study:

• To introduce and investigate a novel concept called running entropy mapping.
• To demonstrate how this method can overcome limitations in detecting causal relationships in nonlinear systems.

Main Methods:

• Development of running entropy mapping to transform time series.
• Application of linear parametric time series methods, such as partial directed coherence, to the transformed sequences.

Main Results:

• Running entropy mapping successfully converts nonlinear coupling problems into a format amenable to linear analysis.
• The method exposes causal relationships that were undetectable using traditional linear approaches.

Conclusions:

• Running entropy mapping offers a powerful new tool for causal inference in complex, nonlinear time series.
• This approach simplifies the analysis of nonlinear systems, making previously hidden causal links discoverable.