舆论动力学中的反转演化机制研究

发布时间：2018-10-12 09:37

【摘要】：随手机等移动终端设备的普及,基于移动互联网的社交平台正在逐渐替代传统平面媒体在舆论导向方面的作用。移动互联网络所特有的移动性、实时性、自媒体特性,使得网络舆论可以快速、广泛的传播,也正由于以上特点,基于移动网络的网络社交平台中舆情反转事件逐年增高,由此产生的网络暴力层出不穷,给政府相关部门的监管带来较大挑战。本文在大量研究国内外舆论动力学与复杂网络方面的文献的基础上,建立了适应移动社交环境的舆论动力学模型,并从仿真层面分析了舆论反转的形成原因,最后编写了舆论倾向性分析平台,本文主要成果如下:1.总结了国内外学者对复杂网络建模与舆论动力学模型的概况,并且针对舆论动力学有限信任模型进行了详细介绍与分析。结合移动互联网社交平台观点交互现状,总结原HK模型缺点,给出改进方向。2.依据移动互联网社交平台无连边个体影响情况,提出基于虚拟引力的舆论动力学模型。并且考虑用户之间网络连接,将发现的互联网平台观点交互自然规律融入其中。体现了真实移动互联观点交互网络中的特性,社会人群不是互相隔离的,每个个体间都存在虚拟的联系,非连接的节点之间也会产生观点的共鸣。3.在基于虚拟引力的舆论动力学模型基础上,引入BA无标度网络模拟真实移动互联网中社交平台用户观点交互网络,通过不同参数组合研究舆论反转现象的成因,总结极端观点个体数量对群体观点演化的影响规律,对政府机构在舆论导向的控制上有一定的现实意义。并且使用真实网络数据进行模型的验证,证明本文提出的模型在一定程度上符合真实舆论反转事件。4.搭建舆论倾向性分析平台。利用基于Python2.7的Scrapy框架编写适用于新浪微博的网络爬虫程序,递归地爬取指定ID下粉丝与其关注对象的用户信息和评论数据,并搭建真实微博用户网络模型。使用Matlab搭建舆论动力学仿真环境,嵌入本文提出的基于虚拟力的舆论动力学模型,通过界面化交互平台输入相应数据,可快捷方便的进行仿真,得到观点演化图像,有一定的工程实用性。
[Abstract]:With the popularity of mobile devices such as mobile phones, the social platform based on mobile Internet is gradually replacing the traditional print media in the role of public opinion guidance. Because of the mobility, real-time and self-media characteristics of mobile Internet, public opinion can spread rapidly and widely. Because of the above characteristics, the public opinion reversal events in the mobile network-based social platform are increasing year by year. The resulting network violence emerges in endlessly, which brings great challenge to the supervision of relevant government departments. Based on a large number of literature on public opinion dynamics and complex networks at home and abroad, this paper establishes a public opinion dynamics model adapted to mobile social environment, and analyzes the reasons for the formation of public opinion reversal from the aspect of simulation. Finally, the platform of public opinion tendency analysis is compiled. The main results of this paper are as follows: 1. In this paper, the general situation of complex network modeling and public opinion dynamics model is summarized, and the finite trust model of public opinion dynamics is introduced and analyzed in detail. Combined with the status quo of mobile internet social platform interaction, the shortcomings of the original HK model are summarized, and the direction of improvement is given. 2. According to the influence of individuals without connected edges on mobile Internet social platform, a public opinion dynamics model based on virtual gravity is proposed. Considering the network connection between users, the interactive natural law of Internet platform is incorporated into it. It reflects the characteristics of the real mobile interconnected viewpoint interactive network, social groups are not isolated from each other, each individual has a virtual connection, and disconnected nodes will also have a resonance of views. 3. On the basis of the public opinion dynamics model based on virtual gravity, BA scale-free network is introduced to simulate the interactive network of social platform users in the real mobile Internet, and the cause of public opinion reversal is studied by combining different parameters. Summing up the influence of the individual number of extreme views on the evolution of group views is of practical significance to the control of public opinion orientation by government agencies. The real network data is used to verify the model, which proves that the model proposed in this paper is consistent with the real public opinion reversal event to a certain extent. 4. Build the public opinion tendency analysis platform. The Scrapy framework based on Python2.7 is used to compile the web crawler program for Sina Weibo. The user information and comment data of fans and their objects of concern under specified ID are crawled recursively, and a real user network model of Weibo is built. The simulation environment of public opinion dynamics is built with Matlab, and the model of public opinion dynamics based on virtual force is embedded in this paper. The corresponding data can be input through the interface interactive platform, and the view evolution image can be obtained quickly and conveniently. It has certain engineering practicability.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：G206

【参考文献】