Contagion Source Detection by Maximum Likelihood Estimation and Starlike Graph Approximation

Detecting the origin of an epidemic-like spread in extensive networks has crucial implications, including pinpointing rumor sources in online social networks. When provided with a snapshot of the nodes harboring the harmful information in the network, the challenge lies in identifying the source of the spread. Our approach tackles source detection using maximum likelihood estimation, introducing a message-passing algorithm tailored for large networks. The versatility of our method is showcased through a novel starlike graph approximation, broadening its applicability to diverse types of graphs.