Optimal reconstruction of dynamical systems: A noise amplification approach

TL;DR

This paper introduces a novel cost function based on noise amplification for optimizing the reconstruction of dynamical systems from time series data, enabling better parameter selection and comparison of reconstruction methods.

Contribution

It proposes a new objective function grounded in noise amplification theory to evaluate and optimize state space reconstructions from time series.

Findings

The cost function effectively identifies optimal delay and embedding parameters.

It allows comparison of different reconstruction approaches.

Demonstrated on synthetic and real data.

Abstract

In this work we propose an objective function to guide the search for a state space reconstruction of a dynamical system from a time series of measurements. This statistics can be evaluated on any reconstructed attractor, thereby allowing a direct comparison among different approaches: (uniform or non-uniform) delay vectors, PCA, Legendre coordinates, etc. It can also be used to select the most appropriate parameters of a reconstruction strategy. In the case of delay coordinates this translates into finding the optimal delay time and embedding dimension from the absolute minimum of the advocated cost function. Its definition is based on theoretical arguments on noise amplification, the complexity of the reconstructed attractor and a direct measure of local stretch which constitutes a novel irrelevance measure. The proposed method is demonstrated on synthetic and experimental time series.