微博信息传播网络的属性研究

发布时间：2018-06-09 21:31

本文选题：微博信息传播网络 + 复杂网络　；参考：《太原理工大学》2014年硕士论文

【摘要】：微博是最近几年才发展起来的新兴事物,它是互联网领域的一个里程碑,微博的出现彻底改变了人们的日常行为方式。微博自诞生以来就凭借其操作简单和功能实用等优点迅速吸引了数量庞大的用户群。微博信息传播属性的研究可以从用户关系网络、信息传播网络和传播机制三个方面出发。微博中用户是微博的主体和核心,用户之间的“关注”和“被关注”关系形成了一个关系网络,显然,这个网络是有方向的。网络是信息传播的载体,每一个消息帖子的传播都需要借助该网络。本文首先介绍了微博的一些基本知识,包括微博的概念、国内外的研究现状、发展历程、主要应用和研究关键问题等。微博是一个复杂的系统,包含了意思不同的各种各样的特殊符号。它的复杂性还体现在用户、微博消息和使用动机的多样性,本文对用户和消息进行了分类。信息之所以能够在微博系统迅速传播,主要依靠其自身的传播动力,因此微博传播动力的研究对深层次地揭示微博特征非常的重要。微博用户间的关系组成一个静态网络图,可尝试从对复杂网络研究的角度去研究微博信息传播网络。本文在大量用户关系的数据基础上,构建了一个庞大的网络,利用VC++、Gephi和MATLAB等多种工具分析了网络的度、聚类系数、平均路径长度、K-核心和社区等特征值,验证了微博信息网络具有小世界、无标度的特征,并且其入度和K-核的分布图具有幂律分布的特点,这为后面模型的构建和改进提供了理论依据。在特征值相关性分析中,发现特征向量中心度与聚类系数间的相关系数很大,它们的关联度很高,这都是衡量网络中心性的指标。度与图密度间的关联度也相对较高,而度与特征向量中心度、聚类系数关联度较低,说明如果一个网络的度很大,它的图密度也一般很大,但中心性可能不强,“小世界现象”也就不明显。在信息传播网络上,提出了三种微元结构,通过实验得出信息分散结构的数量最多,并且随着实验信息链的增多,其数量也逐渐增多。 SI模型、SIS模型和SIR模型是疾病传播研究领域经典的模型,曾经为一些流行病的控制起到了决定性的作用。这三种模型主要针对的是生活中疾病的传播,而疾病的传播与微博上的信息传播有所不同,比如疾病的传播受天气、气候等环境因素的影响,而微博信息的传播受用户心情、生活状况的影响。本文构建了一种符合微博信息传播网络的IOSIR模型,该模型考虑到系统的输入和输出情况,通过四组不同数值的模拟仿真,发现该模型能较准确地揭示微博的信息传播过程。
[Abstract]:Micro-blog is a new emerging thing that has developed in recent years. It is a milestone in the Internet field. The appearance of micro-blog has completely changed people's daily behavior. Since its birth, micro-blog has quickly attracted a large number of users with its advantages of simple operation and functional and practical advantages. The research on the information dissemination property of micro-blog can be made. From the three aspects of the user relationship network, the information communication network and the communication mechanism, the users of micro-blog are the main body and core of micro-blog. The "concern" and "concerned" relationship between users form a relational network. Obviously, the network has a direction. The network is the carrier of information dissemination, and the communication of every message post needs to be transmitted. Use the network.
This article first introduces some basic knowledge of micro-blog, including the concept of micro-blog, the current research situation at home and abroad, the development process, the main application and the key problems of research. Micro-blog is a complex system which contains a variety of special symbols with different meanings. Its complexity is also embodied in the variety of users, the messages of micro-blog and the diversity of motivation. This paper classifies users and messages. The reason why information is able to spread rapidly in micro-blog system depends on its own transmission power, so the research on micro-blog's transmission power is very important to reveal the features of micro-blog.
The relationship between micro-blog users consists of a static network graph, which can try to study the micro-blog information communication network from the perspective of complex network research. Based on the data of a large number of user relations, this paper constructs a huge network, and analyzes the degree, clustering coefficient, and average path length of the network by using VC++, Gephi and MATLAB. The eigenvalues of K- core and community verify that the micro-blog information network has the characteristics of a small world and no scaling, and its admission and K- kernel distribution maps have the characteristics of power law distribution. This provides a theoretical basis for the construction and improvement of the latter model. In the correlation analysis of eigenvalues, the correlation between the center degree of the eigenvector and the clustering coefficient is found. The coefficient is very high, and the degree of association is very high. This is the index of the centrality of the network. The degree of correlation between the degree and the graph density is also relatively high. The degree and the center degree of the eigenvector and the clustering coefficient are relatively low. It shows that if the degree of a network is very large, the density of the graph is also very large, but the centrality may not be strong, "small world phenomenon". It is not obvious. In the information communication network, three kinds of microelement structures are proposed. The number of information dispersing structures is the most in the experiment, and the number of the information chain is increasing with the increase of the experimental information chain.
SI model, SIS model and SIR model are classic models in the field of disease transmission research, which have played a decisive role in the control of some epidemics. These three models mainly focus on the spread of diseases in life, and the spread of diseases is different from that on micro-blog. For example, the spread of diseases is affected by weather, climate and other environmental factors. The transmission of micro-blog information is affected by the mood of the users and the influence of the living conditions. This paper constructs a IOSIR model which conforms to the micro-blog information communication network. The model takes into account the input and output of the system, and finds that the model can reveal the information dissemination process of micro-blog more accurately through four different numerical simulation simulations.
【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.092

【参考文献】