PRIA: a Multi-source Recognition Method Based on Partial Observation in SIR Model

Abstract

Nowadays, the spread of Internet rumors and viruses has caused great hidden dangers to the safety of human life. It is particularly important to identify the source of network threat, especially when there are multiple sources in the network. At present, the research on multi-source propagation is mostly based on SI model, but there is little work on multi-source propagation under SIR model. Based on SIR propagation model, this paper proposes a novel PRIA algorithm to locate multiple propagation sources. Firstly, we propose a new partitioning method based on effective distance, which transforms the source problem into a single source problem in multiple partitions. Secondly, we propose a single source algorithm based on SIR propagation model, which uses reverse infection algorithm to locate suspicious sources. Finally, we evaluate our approach in real network topology. The simulation results show that our method can effectively identify the real source and estimate the propagation time. And it has great accuracy in the number of identification sources.