A Simple Method for Locating Topic Sources in Uncertainty Diffusion Networks

Abstract

The online social network has become an important information transmitting media. Finding the sources of topic on the online social networks (OSNs), which helps us to know the cause of events, to identify the authenticity of the topic and to target the origin of rumors, is very important. However, topic sources locating is difficult because the structure of OSNs is complex and the complete OSNs are hard to observe. At the same time, the fact that the topic diffusion process is multiple-source, asynchronous and uncertain, causes topic sources locating to be more difficult. In this paper we aim to solve the problem of the topic sources locating in the uncertainty diffusion networks. We first analyzed the uncertainty diffusion relationship in detail through the real data sets. Then according to the characteristics of incomplete observation and asynchronous transmission, we propose a dynamic locating method based on the activation time and the partial topology structure from a subnet transition to earlier subnet. Our method needn't build the underlying network topology in advance. Experimental results on BA networks show that our method can solve the topic sources locating problem of asynchronous information diffusion with incomplete observation.