Impact of node dynamical parameters on structures identification of complex networks based on the Lasso method

Complex networks are ubiquitous in nature and society. The functions and features of complex networks are various when these networks have different nodal dynamics and network topologies. Reconstructing networks with high-order nodal dynamics or different system parameter vectors from limited measurable information is a fundamental problem for using and controlling these networks. Based on the Lasso method, we present an efficient and feasible, completely data-driven approach to predict the structures of complex networks in the presence or absence of noise when the systemic parameter is uncertain, that is, the node dynamical parameter vector of network can vary. The numerical simulations indicate that, networks structures can be fully reconstructed even only few information available under the conditions of the systemic parameter vector is varying and in the presence or absence of noise, this method is effective and robust.