一种改进的SEIR网络谣言传播模型研究  

Research on an Improved SEIR Network Rumor Propagation Model

作　　者：范纯龙[1] 宋会敏[1] 丁国辉[1] 

机构地区：[1]沈阳航空航天大学计算机学院,沈阳110136

出　　处：《情报杂志》2017年第3期86-91,共6页Journal of Intelligence

基　　金：国家自然科学基金青年项目"基于查询日志的数据库自动模式匹配技术研究"(编号:61303016)研究成果之一

摘　　要：[目的/意义]在线社交网络的流行推动了网络舆情的研究,网络谣言作为网络舆情分析的重要方面,对其传播规律和控制方法的研究越发重要。[方法/过程]当前,利用传染病模型分析网络谣言传播的多种模型中,SEIR模型通过节点的易感、潜伏、感染、免疫四种状态较全面的反应了谣言的传播特性。但模型对于单位时间内网络节点间的传播概率均等这一假设与实际社交网络的传播性质不一致,个体间的亲密度、谣言接收次数等均影响谣言的传播概率,另外免疫节点对谣言传播在实际情况中是有一定抑制作用的,基于这两点提出一种改进的谣言传播模型PSEIR。[结果/结论]将该模型在无标度网络上进行谣言传播过程仿真,分析影响谣言传播的主要因素,并与传统的SEIR、SIR谣言传播模型进行对比,结果表明PSEIR模型能更好地反映社交网络中谣言的传播规律。[ Purpose/Significance] The popularity of online social network has promoted the research of network public opinion. As an important aspect of network public opinion analysis, network rumor research of its propagation law and control method is becoming more and more important. [ Method/Process] At present, as one of the many models using the epidemic model to analyze the network rumors, SEIR model comprehensively reflects the propagation characteristics of rumors through the four states of the node＇s susceptibility, latency, infection and immunity. But the assumption of the equality of propagation probability among network nodes in a unit time of the model is not consistent with the actual social network and the degree of intimacy between individuals, the times of rumor reception will have impact on the propagation probability of a rumor. In addition, the immune node has a certain inhibitory effect on the spread of rumors in the actual situation. An improved model of rumor propagation PSEIR is then proposed based on the previously mentioned two points. [ Result/Conclusion] The model is simulated on the scale free network and the main factors influencing the spread of rumors are analyzed. And comparison with the SEIR, SIR rumor propagation models shows that the PSEIR model can better reflect the propagation of rumors in social networks.

关 键 词：谣言传播 社交网络 传染病模型 传播概率 无标度网络 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]