基于自适应策略的网络舆论演化  

Network public opinion evolution based on adaptive strategy

作　　者：吴越[1] 

机构地区：[1]西华大学计算机与软件学院,成都610039

出　　处：《计算机应用》2017年第A01期283-286,共4页journal of Computer Applications

基　　金：国家自然科学基金资助项目(61602389);四川省教育厅科研项目(15ZB0133);西华大学自然科学重点基金项目(z1422617);西华大学省部级学科平台开放课题(szjj2015-58);互联网自然语言智能处理四川省高等学校重点实验室资助项目(INLP201501)

摘　　要：社交网络上的舆论演化过程是现今国内外信息传播、复杂系统、行为分析等领域的一个研究热点。针对现有研究常受制于固定的网民观点交换、更新策略,而难以还原真实网络舆论演化过程的问题,提出了一种基于自适应策略的网络舆论演化模型。首先,选取BA(Barabási和Albert)无标度网络和WS(Watts和Strogatz)小世界网络作为网络模型。然后,结合Sznajd模型与多数准则模型的思想,设计了两种观点更新策略。最后,为保证网民所持观点在全局网络中占据优势,设置了自适应的观点更新策略选取规则。仿真实验结果表明,随着舆论演化时间的增长,网民选取固定策略的概率逐渐趋近于1;最终选择说服邻居策略的网民数量与选择多数观点策略的网民数量基本相等,这种变化趋势不依赖于初始观点分布和网络拓扑结构。在舆论演化过程中,节点中心度大于100的网民均有固定的观点更新策略,而中心度小于10的网民常动态改变策略。仿真实验结果显示,WS网络和BA网络中持少数观点的网民数量会随着时间增长而减少,但WS网络中的少数观点网民数量减少幅度更大。The study of the public opinion evolution on social networks is a hotspot across the world recently, in the fields of information transmission, complex systems and behavior analysis. The existing research is often limited by the fixed renewal and switching strategies of users＇views, and difficult to restore the real evolution process of the network public opinion. In order to solve the problem, a network public opinion evolution model based on the adaptive strategy was proposed. Firstly, the BA（ Barabási and Albert） scale-free network and WS（ Watts and Strogatz） small-world network were selected as the network models. Secondly, combined with the idea of Sznajd model and majority rule model, two opinion updating strategies were designed. Lastly, for occupying the advantage in the global network, adaptive rules to select the updating opinion strategies were proposed. The experimental results show that, with the development of public opinion, the probability of choosing fixed strategy is gradually approaching 1. In the end of the evolution, the number of netizens who choose to persuade their neighbors will be equal with that of users who select the strategy of majority views, the result is not dependent on the initial distribution of views and the network topology. In the evolution process of public opinion, the users whose degrees are greater than 100 use a fixed strategy, however, the users whose degrees are smaller than 10 always vary their strategies dynamically. Simulation results also show that, the number of users with minority views will be reduced in both WS and BA networks, and the number fells more steeply in the WS network.

关 键 词：社交网络 舆论演化 自适应策略 拓扑结构 节点中心度 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]