Locating the propagation source in complex networks with a direction-induced search based Gaussian estimator

Abstract

Locating the propagation source is crucial for developing strategies to control the spreading process taking on complex networks. Gaussian estimator (GE) is one of the most effective methods for propagation source locating in networks with limited observers. However, on general graphs, due to GE makes an approximation that the actual diffusion tree in spreading process is assumed to be a breadth-first search (BFS) tree, and thus ignores the effect of direction information recorded in observers. Therefore, the accuracy of GE is affected. In this paper, by utilizing the direction information in observers, we define, for the first time, a novel direction-induced search (DIS). Further, a direction-induced search based Gaussian estimator (DISGE) is proposed by combining DIS and the original GE. Experimental results on a series of synthetic and real networks show that the DISGE is feasible and effective in locating the propagation source with limited observers.