Network reconstruction by linear dynamics

Inferring connectivity of complex network from its node dynamics is of great significance in many cross-disciplinary fields. In this paper, we extend the previous network reconstruction method based on phase synchronization with nonlinearly coupled system to a linearly coupled one. We test these two methods and prove that both of them are feasible for the reconstruction of any unknown network structure. Compared to the nonlinear method, however, we find that from the computational efficiency point of view the linear reconstruction method is always superb, showing its own unique merits, such as simpler in the algorithm realization, faster in the computational time, and more efficient in the final reconstruction result. On the other hand, we also briefly discuss the limitation of the linear reconstruction method.