Derivative-variable correlation reveals the structure of dynamical networks

TL;DR

This paper introduces a novel method for reconstructing the topology of dynamical networks by analyzing correlations between node variables and their derivatives, applicable to weighted or directed networks with known interaction functions.

Contribution

The paper presents a new matrix-based approach using variable-derivative correlations for network reconstruction, with an adjustable parameter to optimize accuracy.

Findings

Method successfully reconstructs network topology in example scenarios.

Reconstruction accuracy depends on the dynamical properties of the time series.

Applicable to any weighted or directed network with known interaction functions.

Abstract

We propose a conceptually novel method of reconstructing the topology of dynamical networks. By examining the correlation between the variable of one node and the derivative of another node, we derive a simple matrix equation yielding the network adjacency matrix. Our assumptions are the possession of time series describing the network dynamics, and the precise knowledge of the interaction functions. Our method involves a tunable parameter, allowing for the reconstruction precision to be optimized within the constraints of given dynamical data. The method is illustrated on a simple example, and the dependence of the reconstruction precision on the dynamical properties of time series is discussed. Our theory is in principle applicable to any weighted or directed network whose internal interaction functions are known.