【作者】 孙海燕；

【导师】 江铭炎； 汤代禄；

【作者基本信息】 山东大学 ， 电子与通信工程（专业学位）， 2014， 硕士

【摘要】 近年来,网络技术得到了快速发展,各种网络应用平台随之进入人们的视野,给人们的工作、学习和生活带来了极大的便利,比如,微博、博客、QQ空间、网络社区等。这些网络应用平台给信息的共享与传播提供了新的方式,加快了信息的传播速度,缩短了人们之间的距离。在这些网络应用平台中,每一个人都可以成为信息的发布者与转载者,彻底颠覆了传统的新闻传播模式。新的信息发布与传播模式是把双刃剑,既有有利的一面,也不乏危害之处。一方面,快速的信息传播模式促进了科技的发展,提高了劳动生产率,推动了人类历史的进步；另一方面,它也给谣言的传播创造了良好的条件,使得谣言传播更加便捷、更加隐蔽。社会生活中的热点事件经过网络的传播、放大、甚至是扭曲,往往就会在极短的时间内形成一股强大的舆论力量。由于网络用户规模的庞大性,网络舆情在一定程度上代表了社会舆情,对国家政策的制定、是非判断都产生重要的影响。因此,国内外都对网络舆情研究非常重视。但是目前的研究缺少理论与实际案例的结合,缺少社会科学与自然科学的交叉研究。首先,从实际的网络舆情案例开始着手,用统计学的方法,分析了2013年初发生的袁厉害弃婴收容所火灾事件所引起的网络舆情。以此案例为基础,把整个网络舆情传播过程划分为五个阶段,即：发生阶段、爆发阶段、过度阶段、蔓延增长阶段和衰退阶段。然后,进一步用此案例对现有的典型网络舆情传播模型进行了验证。研究结果表明原模型基本上能够反映网络舆情发展过程中,关注人数的总体演化趋势,但是没有能够较好地反映关注人数的动态变化情况。根据人们的直觉,在网络舆情发展过程中,对于网络热点事件的关注人数时多时少,但是总体趋势是先多后少,然后逐渐消亡。为此,本文进一步对网络舆情传播模型进行了改进,提出了带有变量参数的网络舆情传播模型。研究结果表明,采用变量动态参数,使网络舆情传播模型不仅能够反映网络舆情发展过程中参与人数的总体演化趋势,而且较好地解释了网络舆情发展过程中参与人数的动态变化。数据仿真结果显示,在舆情的发生及爆发阶段参与率大于等于退化率,且在其它阶段保持参与率小于等于退化率。其次,本文对随机网络中的舆情传播进行了深入的研究,提出了随机网络中的舆情传播模型。研究结果表明,随机网络中的舆情传播速度与随机网络节点的密度、节点的信号辐射半径、节点的移动速度、信息传播率以及信息的退化率等参数密切相关。只要随机网络节点保持以v=2.0km/m的速度运动在节点密度为40个/平方千米的区域中并且能探测到半径为0.1km的范围内的节点,以传播率α=1.0和衰减率β≤0.16时,网络舆情就够在整个网络上快速传播。更多还原

【Abstract】 In recent years, the network technology has been well developed. A variety of network application platforms has come into our view. It brings great convenience in our work, study and life. For example, micro-blog, blog, QQ space, online communities and so on. These network application platforms provide a new way for information sharing and dissemination. They accelerate the speed of information dissemination, shorten the distance between people. In these network application platforms, everyone can be a publisher and forwarder. It revolutionizes the traditional news media modes. The new mode of information dissemination and communication is a double edged sword. There are positive sides and some of the hazards. On the one hand, the rapid information dissemination model promotes the development of technology, improve labor productivity, and promote the progress of human history. On the other hand, it also creates good conditions for the dissemination of rumors, making the rumors spread more convenient, more subtle. Once the social life hot events have been spread enlarged or distorted over the network, they tend to form a powerful force of public opinion in a very short time. Due to the huge size of network users, network public opinion, to some extent, represents the public opinion of the community. It has an important impact on the formulation of national policy, non-judgmental. Therefore, the public opinion research has been gotten a very important attention all over the world. However, current researches are lacks of both combination of theory and practical study cases and the cross-over of the social sciences and natural sciences.First, we begin our research from a practical case of network public opinion. We analyze fire incident of the abandoned shelters of Yuan Lihai, occurred in early2013. Based on public opinion cases, we divide the whole network opinion dissemination process into five stages, namely the occurrence phase, outbreak phase, the transitional phase, spreading growth phase and decline phase. Then, with this case, the existing network of public opinion typical propagation model is validated. The results shows that the original model can basically reflect the overall evolution trend of the number of people concerned about the issue, but there is no better able to reflect the dynamic changes in the number of concerns. According to our intuition, in the whole public opinion development process, the number of people concerned about the hot event is fluctuated, but the overall trend is increasing at first, then, decrease gradually. To this end, further propagation model for network public opinion has been proposed to improve the network public opinion propagation model with variable parameters. The results show that the variable dynamic parameters can not only reflect public opinion propagation in the number of participants in the development process of the evolution of the overall trend, but has a better explanation of the dynamic changes in the network of public opinion numbers involved in the development process. Data Simulation results show that in the incidence of public opinion and the outbreak stage the participation rate greater than or equal the degradation rate and the participation rate in the other stages remain less or equal than the degradation rate.Secondly, the spread of public opinion in random networks is conducted in-depth research. A dissemination of public opinion model in random networks is proposed. The results show that the density of a random network, random network node, the signal radius of the node, the node moving speed, the information transmission rate and degradation rate information parameters are closely related to the propagation velocity of the public opinion. If the random network node holds v=2.0km/m and the network keeps node density in40/sq km, the node radius is able to detect the nodes within the range of0.1km, the propagation rate α=1.0and the decay rate is β≤0.16, public opinions can be spread in the random networks.更多还原