基于超网络的网络舆情分析研究与应用
发布时间：2019-06-14 19:56
【摘要】：随着信息化进程的深入,网络成为继报纸、广播和电视等传统媒体之后的“第四媒体”,对人们工作、生活和社会舆论格局的影响日益深入,成为社会民意的晴雨表和民众表达诉求的重要平台。然而,网络在降低民众言论表达门槛的同时,也成为谣言滋生、扩散的重灾区。网络舆情在产生、传播的过程中受到来自多个方面因素的影响,选取合适的模型对网络舆情进行分析研究显得日益重要。本文旨在结合超网络理论对网络舆情分析与应用进行研究,在网络舆情分析建模的基础上探索合适的网络舆情引导策略。具体内容包括如下三个方面:构建基于超网络的网络舆情分析框架。在分析网络舆情产生与演化机制及影响因素的基础上,将网络舆情的演化过程抽象为舆情主体在环境外驱力和心理内驱力的共同作用下发表舆论观点。从中提取舆情主体节点、驱动力节点和舆论观点节点三类节点,将节点中同质节点间连接构成子网络层,异质节点间连接构成超边,构建了一个三层的超网络舆情分析模型,为网络舆情分析提供基本研究框架。同时,对网络舆情分析模型中的各类节点、子网络层和超边的属性进行阐述,并通过类比社会网络分析的相关概念选取超网络的测度指标。构建基于超网络的关键因素识别模型。通过子网络层分析、超边排序、关键因素识别三个步骤构建基于超网络的网络舆情关键因素识别模型。在子网络层分析过程中从广度和深度两方面对环境子网进行分析,结合犹豫模糊语言术语集对心理子网进行分析,结合向量空间模型对观点子网进行分析;在此基础上结合超诱导主题搜索算法提出改进的超边排序算法并构建关键因素识别模型,进而识别出各子网络层的关键因素:舆情领袖、关键环境信息、主流心理情况和主流舆论观点。探索基于超网络的网络舆情分析的应用。通过运用算例分析和综合评价方法验证基于超网络的网络舆情关键因素识别模型的合理性和可靠性。并根据模型分析的结果,探索基于超网络的网络舆情分析的应用点和基于超网络的网络舆情引导策略。
[Abstract]:With the deepening of information process, the network has become the "fourth media" after newspapers, radio and television and other traditional media, which has a deeper and deeper influence on people's work, life and public opinion pattern, and has become an important barometer of social public opinion and an important platform for the people to express their aspirations. However, while lowering the threshold for public expression, the Internet has also become a disaster area where rumors breed and spread. The network public opinion is influenced by many factors in the process of dissemination, so it is more and more important to select the appropriate model to analyze and study the network public opinion. The purpose of this paper is to study the analysis and application of network public opinion based on hypernetwork theory, and to explore the appropriate guiding strategy of network public opinion on the basis of network public opinion analysis and modeling. The specific content includes the following three aspects: the construction of network public opinion analysis framework based on hypernetwork. Based on the analysis of the mechanism and influencing factors of the emergence and evolution of network public opinion, the evolution process of network public opinion is abstracted into the expression of public opinion by the subject of public opinion under the joint action of external driving force and psychological driving force. Three kinds of nodes, namely, the main node of public opinion, the driving node and the node of opinion of public opinion, are extracted, and the connection between the homogeneous nodes in the node is formed into a sub-network layer, and the connection between heterogeneous nodes is formed into a superedge. A three-layer hypernetwork public opinion analysis model is constructed, which provides a basic research framework for the analysis of network public opinion. At the same time, the attributes of all kinds of nodes, subnetwork layer and superedge in the network public opinion analysis model are described, and the measurement index of hypernetwork is selected by analogy with the related concepts of social network analysis. The identification model of key factors based on hypernetwork is constructed. Through the analysis of subnetwork layer, hyperedge sorting and key factor identification, the recognition model of key factors of network public opinion based on hypernetwork is constructed. In the process of subnetwork layer analysis, the environmental subnet is analyzed from two aspects of breadth and depth, the psychological subnet is analyzed with hesitant fuzzy language terminology set, and the viewpoint subnet is analyzed with vector space model. On this basis, an improved hyperedge sorting algorithm is proposed and a key factor recognition model is constructed, and then the key factors of each sub-network layer are identified: public opinion leader, key environmental information, mainstream psychological situation and mainstream public opinion. Explore the application of network public opinion analysis based on hypernetwork. The rationality and reliability of the recognition model of key factors of network public opinion based on hypernetwork are verified by example analysis and comprehensive evaluation method. According to the results of model analysis, the application point of network public opinion analysis based on hypernetwork and the guiding strategy of network public opinion based on hypernetwork are explored.
【学位授予单位】：南京师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：C912.63
[Abstract]:With the deepening of information process, the network has become the "fourth media" after newspapers, radio and television and other traditional media, which has a deeper and deeper influence on people's work, life and public opinion pattern, and has become an important barometer of social public opinion and an important platform for the people to express their aspirations. However, while lowering the threshold for public expression, the Internet has also become a disaster area where rumors breed and spread. The network public opinion is influenced by many factors in the process of dissemination, so it is more and more important to select the appropriate model to analyze and study the network public opinion. The purpose of this paper is to study the analysis and application of network public opinion based on hypernetwork theory, and to explore the appropriate guiding strategy of network public opinion on the basis of network public opinion analysis and modeling. The specific content includes the following three aspects: the construction of network public opinion analysis framework based on hypernetwork. Based on the analysis of the mechanism and influencing factors of the emergence and evolution of network public opinion, the evolution process of network public opinion is abstracted into the expression of public opinion by the subject of public opinion under the joint action of external driving force and psychological driving force. Three kinds of nodes, namely, the main node of public opinion, the driving node and the node of opinion of public opinion, are extracted, and the connection between the homogeneous nodes in the node is formed into a sub-network layer, and the connection between heterogeneous nodes is formed into a superedge. A three-layer hypernetwork public opinion analysis model is constructed, which provides a basic research framework for the analysis of network public opinion. At the same time, the attributes of all kinds of nodes, subnetwork layer and superedge in the network public opinion analysis model are described, and the measurement index of hypernetwork is selected by analogy with the related concepts of social network analysis. The identification model of key factors based on hypernetwork is constructed. Through the analysis of subnetwork layer, hyperedge sorting and key factor identification, the recognition model of key factors of network public opinion based on hypernetwork is constructed. In the process of subnetwork layer analysis, the environmental subnet is analyzed from two aspects of breadth and depth, the psychological subnet is analyzed with hesitant fuzzy language terminology set, and the viewpoint subnet is analyzed with vector space model. On this basis, an improved hyperedge sorting algorithm is proposed and a key factor recognition model is constructed, and then the key factors of each sub-network layer are identified: public opinion leader, key environmental information, mainstream psychological situation and mainstream public opinion. Explore the application of network public opinion analysis based on hypernetwork. The rationality and reliability of the recognition model of key factors of network public opinion based on hypernetwork are verified by example analysis and comprehensive evaluation method. According to the results of model analysis, the application point of network public opinion analysis based on hypernetwork and the guiding strategy of network public opinion based on hypernetwork are explored.
【学位授予单位】：南京师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：C912.63
【参考文献】
相关期刊论文 前10条
1 游丹丹;陈福集;;我国网络舆情预测研究综述[J];情报科学;2016年12期
2 张伟;;可视化分析技术在网络舆情研究中的应用[J];现代情报;2016年11期
3 吴树芳;徐建民;;基于HITS算法的微博用户可信度评估[J];山东大学学报(工学版);2016年05期
4 郭秋萍;华康民;;超网络研究分析综述[J];管理工程师;2016年04期
5 杨欣蓉;钱钢;冯向前;;基于犹豫模糊语言多属性群决策的VIKOR扩展方法[J];计算机工程与应用;2017年11期
6 王莹;于海;朱志良;;基于软件节点重要性的集成测试序列生成方法[J];计算机研究与发展;2016年03期
7 冯向前;谭倩云;钱钢;;犹豫模糊语言的可能度排序方法[J];控制与决策;2016年04期
8 黄微;李瑞;孟佳林;;大数据环境下多媒体网络舆情传播要素及运行机理研究[J];图书情报工作;2015年21期
9 马宁;刘怡君;;基于超网络的舆情演化多主体建模[J];系统管理学报;2015年06期
10 胡秀丽;;基于VSM和LDA模型相结合的微博话题漂移检测[J];兰州理工大学学报;2015年05期
相关博士学位论文 前2条
1 陈t熀，
本文编号：2499642
本文链接：https://www.wllwen.com/shekelunwen/shgj/2499642.html
