Analysis Of Information Spreading Path In Social Networks

Spreading phenomena,such as computer virus diffusion,information spreading in social networks and advertising propagation have attracted considerable attention in recent years.Previous works mostly focus on the theoretical analysis and experiment simulation,lacking in-depth investigation of empirical illustrations.Differently,this paper utilizes systematic theory and data mining theory to analyze the empirical spreading paths.The main contribution of the paper includes two parts:The first part mainly focuses on the overlapping influence between multiple spreaders,which provides a forward step to control the spread of information in complex networks.Traditionally,most work selects spreaders depending on the centralities of nodes,such as degree centrality,betweenness centrality and so on,ignoring the coupling effects between spreaders.However,based on the hypothesis of overlapping influence(coupling effects),we proposed a novel framework to upgrade the collective influence of multiple spreaders.Comparing with state of the art methods,our method could select central,yet sparse spreaders,with low overlapping influences.Finally,we simulate spreading paths using SIR model in real networks,which illustrates the effectiveness of our method.The second part focuses on the empirical spreading paths in real location-based networks.We firstly track the spreading paths and analyze the evolving patterns of epidemics.It shows that the spreading probability of information changes with time,which violates the classical spreading models with a constant spreading probability.Moreover,we find a counterintuitive phenomenon that a delay gap exists between the maximal spreading probability and spreading velocity.Then we propose an improved SI spreading model and thoroughly analyze the delay gap.Finally,experiments in artificial networks,including BA and ER networks,illustrate the validity of our model.