Detecting essential nodes in complex networks from measured noisy time series

摘要

A nonlinear measure, namely multi-interdependency, is proposed to detect the essential nodes in heterogeneously dynamical networks. The method is based upon the conceptions of the nearest conditional neighbors and singular value decomposition (SVD). Numerical results show that the value of multi-interdependency is positively correlated with the degree of nodes, which is beneficial to identify the nodes of topological and functional importance. Moreover, such a method has been demonstrated being robust against the effect of intrinsic noise.