Research On Associated Event Dissemination Rule Based On Dual Networks

With the development of society,communication means and communication methods are increasingly enriched with the support of the development of network and communication technologies,and people?s exchanges are becoming more frequent and more convenient.On this basis,social public opinion is accompanied by the propagation of various events to achieve explosive growth.In these incidents,it is very important and common that a type of related incidents cannot be ignored in their interaction dissemination.No matter whether it is a benign event or a malignant incident,it will be fueled by its associated events,which will lead to a stronger social effect.Good use of the rules of communication of associated events can strengthen the degree of popularization of government decisions,help promote corporate culture and brand communication,and effectively promote the development of self-media.Criminals may also use associated events with enhanced effects to spread rumors and anti-social speech.The government can curb the spread of illegal information by breaking the law of transmission of associated events.No matter whether it is the government,enterprises or individuals,it should pay high attention to the implicit dissemination of associated events.Therefore,the research on the propagation law of associated events has great theoretical and practical significance.Through combing the existing papers in the field of information dissemination,it is found that the existing information dissemination model is overly dependent on the epidemic model,and most of the models are improved infectious disease models.The existing research also lacks the analysis of the propagation laws of associated events such as derivative events,and mostly stays in the qualitative perspective or simple quantitative analysis by adjusting the parameters of the infectious disease model.It lacks systematic and comprehensive simulation quantitative analysis.Therefore,based on the percolation theory that can better describe the network structure and the effect of communication,this paper constructs a single-layer network and a dual-network information dissemination model based on self-consistent equations.The main contents of the study include the following three parts:(1)Firstly,on a single-layer network,a percolation model is constructed,which uses a self-consistent equation as a framework to traverse all nodes.This model has higher sensitivity compared to existing percolation models and it can accurately express the phase transition of mutually connected giant component with the probability of percolation.Based on the newly constructed percolation model,the metrics of network diffusion capacity were added to establish a single-layer network information dissemination model.Using this model,the simulation experiments were performed on single-layer networks with different clustering coefficients and variances with different degrees of distribution.It was found that the clustering coefficient was positively correlated with the network's diffusion ability and the variance of the degree distribution was negatively correlated with the network's diffusion ability.(2)In this paper,the association events are defined,and the qualitative and quantitative identification of the event correlations are given.The improved maximum entropy model is compared with other commonly used incident event recognition methods to identify the effect of correlation events.It is found that the improved maximum entropy model has better recognition performance than Bayesian model and K-means model,slightly better than BP neural network.When the extracted feature words are less than or equal to 25(up to 35),it is slightly better than the SVM(support vector machine)model,but the improved maximum entropy model is slightly inferior to the SVM model when the feature words are between 30 and 35.On this basis,the paper further analyzes the promoting effect of the related events,and introduces the promotion factor to redefine the propagation probability of the associated events.(3)Finally,this paper adopts the construction idea of single-layer network information propagation model,and further constructs a dual-network-based association event propagation model based on self-consistent equations.This model is used to further verify and analyze the promotion effect of the association event propagation.Through simulation experiments,it is found that when the promotion factor ? ?,0.3,0.8-,under the effect of the promotion effect,the effect of the network in enhancing the events diffusion is significant.By selecting the incremental Pearson correlation coefficients of the different layer clustering coefficients on the dual network structure,it is found through simulation experiments that the network's diffusion ability is related to the correlation coefficients are positively correlated when the Pearson correlation coefficient of different layer clustering coefficients is(0,1].When the Pearson correlation coefficients of different layer clustering coefficient is [-1,0),the network diffusion capacity is negatively correlated with the correlation coefficient.