网络舆情的信息情感维度空间构建和信息情感元识别研究

发布时间：2018-01-21 09:07

  本文关键词： 网络舆情 信息情感维度 关联模型构建 信息情感空间 信息情感元 情感识别　出处：《吉林大学》2017年博士论文　论文类型：学位论文

【摘要】：在现今移动互联网和大数据的时代背景推动下,网络舆情在表达社情民意、民众心声方面发挥了前所未有的重要作用,舆情信息中蕴含的情感要素作为网络舆情六大要素之一,在社会舆情空间内呈现出更强的巨量性、集聚性、突变性和复杂性,是网络舆情发生发展过程的催化剂。党的十八届三中全会中提出健全坚持正确舆论导向和网络突发事件处置机制,建立良好的网络舆论工作格局。其中及时、精准、高效地把握网络舆情参与民众的情感变化趋向和规律是网络舆情工作格局建设的内容之一,具有重要的理论和实践意义。网络舆情信息情感问题是信息学、心理学和计算机科学相融合的综合性研究,目前已得到国内外相关研究学者的普遍关注,探究网络舆情与信息情感相关的一系列研究问题成为本交叉领域的研究热点,相关研究者也在舆情信息的传播、监测和危机管理等方面做了较多的探讨。实际上网络环境下舆情信息情感具有自身多维交互的动力作用空间,其背后隐藏着诸多极其敏感且可预测的情感变化规律,然而,目前研究成果中对于网络舆情信息情感维度的分析研究尚未涉及,也未将网络舆情信息情感问题放在具有维度结构的完整空间内进行系统建构和关联解析,这也正是网络舆情信息分析与监测研究中亟待重点突破的研究课题之一。本研究带着深入探寻网络舆情信息情感维度空间规律的目的,以情报心理学、情感维度理论、情感计算和情感识别为理论和技术支撑,全面分析网络舆情信息情感维度的内涵、特征、度量、关联作用等,在此基础上探讨了网络舆情信息情感维度空间的建立及子空间作用机制,对于网络舆情信息情感维度空间中基本单位——信息情感元的概念、分布和语义库构成等做了较全面的解析,对网络舆情信息情感元识别目标、要素、模型进行了阐释,并对论文提出的主要模型做了实证研究的尝试,期望进一步表达和量化网络舆情信息情感维度空间特征和演化规律,为网络舆情情感计算领域研究提供基础理论和技术参考。第一章绪论,主要阐述网络舆情信息发展的时代和历史背景、阐述了本研究课题的目的和意义,综述国内外网络舆情信息情感研究的现状,并发现其中存在的关键问题作为本研究的基本出发点,分析研究中使用的方法和技术路线,最后概括了本文的主要研究内容及研究创新点。第二章相关理论综述,阐述了信息论与全信息理论、心理学与情报心理学相关的基本情绪论、离散情感理论、情感维度理论、情感原型理论等几种主要的研究理论;基于以上理论进一步将网络舆情、信息情感和维度空间概念做了综述和界定,同时对网络舆情信息分析的度量方法做了分析和比较,最后对情感计算和情感识别理论和技术的发展研究以及在本研究的应用途径做了阐述。第三章网络舆情信息情感维度分析,依据情感维度相关理论和网络舆情信息情感的特征和属性分析,提出网络舆情的信息情感维度概念,定义了网络舆情情感各特征的内涵、要素,并通过案例分析对网络舆情信息发展中各个维度的特征和表现做了深入探讨,以此作为研究的出发点,为后续关于网络舆情信息情感维度空间的构建做好研究铺垫。第四章网络舆情信息情感维度要素关联模型的构建,首先对网络舆情信息情感维度要素进行提取,对各个维度要素的关联作用进行理论分析,进而构建关联模型,最后验证网络舆情信息情感维度要素关联模型,进一步阐释网络舆情信息情感维度的关联性和演变规律。第五章网络舆情信息情感维度的空间构建,基于前期网络舆情信息情感维度的研究基础,进一步提出了网络舆情信息情感维度空间的概念、构成与特征,提出网络舆情信息情感空间的构建原则,详细阐释了网络舆情信息情感维度空间构建的心理学基础,基于以上分析应用a-v-p心理学模型构建网络舆情信息情感维度空间,进而阐述了情感维度空间中各多维情感综合作用机制,总结了网络舆情信息情感空间运动模式。第六章网络舆情信息情感维度空间中情感元解析,首先阐释网络舆情信息情感元的概念,包括信息情感元的定义、空间特征和分类,剖析网络舆情信息情感元与其所在空间的关联,探讨多维信息情感元语义库的组成方法,探索网络舆情信息情感元在各个子空间中语义鸿沟的解决途径,进而提出信息情感元关联模型、语义判定流程和空间时序特征。第七章基于多维尺度的网络舆情信息情感元的识别模型构建,首先提出网络舆情信息情感元识别的总体框架,包括识别目标、技术方法等,进而提出网络舆情信息情感元关联度量影响模型,对网络舆情信息情感元识别因素进行分析,最后优化了欧式距离的多维尺度识别模型,期望将网络舆情信息情感元放在空间概念下进行识别理论和方法上的尝试。第八章实证研究,根据前期研究基础,对网络舆情信息情感数据进行采集、标注、预处理,然后根据网络舆情信息情感维度空间聚类分析,对空间数据的分布特性、聚类方法、聚类模型进行比较和描述,最后根据前期研究的结果对网络舆情信息情感维度子空间聚类结果、信息情感元识别结果、信息情感维度子空间隶属频率、信息情感元聚集性危机预警结果等进行分析和探讨,进而阐释了信息情感元对舆情演变影响的趋势估计。第九章总结与展望,总结本文的研究工作、取得的创新成果和不足之处,并给出了进一步研究的设想。事实上,网络舆情信息隐含的情感语义内容是具有隐含性、多维性、关联性的一个较为完整的情感空间系统,深入探寻网络舆情信息情感维度空间及信息情感元如何关联和相互作用,有助于深入探寻网络舆情信息中情感数据的内在规律以及多维信息情感数据的作用空间范围,从而为大数据网络舆情的高效引导提供有益参考,提升舆情危机的预测、管理与导向的水平和能力。
[Abstract]:The mobile Internet and big data era to promote in nowadays, network of public opinion in the expression of public opinion, has played an important role in the hitherto unknown voice of the people, public opinion information contained in the emotional factors as one of the six elements in the network of public opinion, public opinion space presents stronger giant, agglomeration, mutation and the complexity is the catalyst development process of network public opinion. In the third Plenary Session of the 18th CPC Central Committee the party puts forward perfect insist on correct guidance of public opinion and network emergency disposal mechanism, establish the network public opinion work pattern is good. The timely, accurate, efficiently grasp the trend and rule of the emotion changes of network public opinion public participation is one of the network public opinion work pattern construction content. It has important theoretical and practical significance. The network public opinion information emotional problem is the integration of information science, computer science and psychology Comprehensive study, has received widespread attention in the related research scholars at home and abroad, to explore a series of problems related to the information of the network public opinion emotion has become a research focus in this cross domain, researchers also in the dissemination of information of public opinion, do more discussion on monitoring and crisis management and other aspects. The dynamic space of public opinion in fact, the emotional information under the network environment has its own multidimensional interaction, hidden behind many extremely sensitive and can predict the emotional changes, however, the current research achievements in the research on information emotional dimension of network public opinion has not yet been involved, nor will the complete spatial information network public opinion on the emotional problems with dimension structure within the system the construction and correlation analysis, which is one of the key breakthrough to the research network public opinion information analysis and monitoring. In this study, with a deep In search of network public opinion information emotion dimension space law, intelligence psychology, emotion theory, affective computing and emotion recognition as the theoretical and technical support, comprehensive analysis of the connotation, characteristics of network public opinion information emotional dimension, measure, correlation function, on the basis of the mechanism and the role of network public opinion information emotion subspace the dimension of space, the concept of network public opinion information in the basic unit -- emotion dimension space information emotional element, distribution and composition of semantic database analysis made more comprehensive, elements of the network public opinion information, emotional element object recognition model are explained, and the main model proposed in the thesis attempts to do empirical research. Expectations of further expression and quantification of network public opinion information emotion dimension space characteristics and evolution rules, to provide the theoretical basis and affective computing research field of network public opinion Technical reference. The first chapter is the introduction, mainly expounds the network public opinion of the development of information era and historical background, expounds the purpose and significance of this research, present situation of domestic and overseas network public opinion information emotion research, and find out the key problems in this study as the basic starting point, methods and techniques used in research and analysis at last, this paper summarized the main research content and innovation of the research. The second chapter reviews relevant theories, expounds the theory of information and the information theory, the basic emotions associated with intelligence psychology theory, discrete emotion theory, emotion theory, emotion theory and several prototype main research theory; based on the above theory to network public opinion, information and emotional dimensions of spatial concepts are discussed and defined, and the method of measurement and analysis of network public opinion information are analyzed and compared, finally. Study on the development of a sense of computing and emotion recognition theory and technology and application methods in this study are described in detail. The third chapter is the analysis of network public opinion information based on the characteristics and attributes of emotional dimension, emotional dimension analysis theory and network public opinion information of emotion, emotional dimension concept put forward the information of network public opinion, defines the connotation of network public opinion, the emotion the characteristics of the elements, and through the case analysis of the features and performance of each dimension in the development of network public opinion information is further analyzed, as the starting point of the research, for the subsequent construction of the network public opinion information emotion dimension space to do research groundwork. The fourth chapter is the structure of the network public opinion information emotion feature relation model, first on the network the emotional dimension elements of public opinion information extraction, theoretical analysis of correlation effect on each dimension elements, then construct the relational model, finally Verify the network public opinion information emotional factors related model, further information the emotional dimension of the relevance and interpretation of network public opinion evolution. The fifth chapter network public opinion information emotion dimension space construction, the former research network public opinion information based on the emotional dimension, further put forward the concept of network public opinion information emotion dimension space, composition and characteristics. The construction principles of network public opinion information emotional space, explains the psychological basis of network public opinion information emotion dimension space construction, the above analysis using a-v-p model of network public opinion information psychology emotion dimension space based on, and expounds the comprehensive function of the multidimensional emotion mechanism of emotional dimensions, summarizes the network public opinion information emotional space movement patterns. Sixth chapter network public opinion information emotion dimension space emotional element analysis, first explains the network public opinion letter The concept of information and emotional element, including the definition of emotional information element, spatial characteristics and classification, correlation analysis of network public opinion information emotional element with the space, explore the method of multidimensional information element emotion semantic database, explore the solution of network public opinion information emotional element in each sub space semantic gap, and then put forward the emotional information element correlation model, semantic feature and spatial temporal decision process. The seventh chapter construct the recognition model of network public opinion information emotional element based on multidimensional scaling, first put forward the overall framework of the network public opinion information emotion recognition, including target identification, technical methods, and then puts forward the network public opinion information emotional element association metric model, the network public opinion information emotion element identification factors were analyzed, and finally optimize multi dimension recognition model of Euclidean distance, the expectation of network public opinion information on the emotional element space Try to identify the concept of theory and method. The eighth chapter is the empirical research, based on previous research, information on emotional data online public opinion collection, annotation, pretreatment, and then according to the network public opinion information emotion dimension spatial clustering analysis, clustering distribution characteristics of spatial data, clustering models are compared and described. Finally, based on the previous research results of network public opinion information emotional dimension subspace clustering results, information element emotion recognition results, information emotion subspace membership information aggregation frequency, emotional element crisis warning results are analyzed and discussed, and then explain the emotional impact on the evolution of public opinion information element trend estimation. The ninth chapter is summary and outlook the research work, summarizes the achievements and deficiencies of innovation, and gives suggestions for further research. In fact, the network public opinion information hidden The semantic content is implicit, multi-dimensional, a more complete system of the relevance of the emotional space, further explore the emotional dimension of network public opinion information space and information element to emotional interrelation and interaction, is helpful to further explore the network of public opinion information according to the numbers of internal rules and multidimensional information data the role of emotion the space scope, so as to provide beneficial reference for efficient data guide network public opinion, enhance the prediction of public opinion crisis, guiding and management level and ability.

【学位授予单位】：吉林大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：C912.63
，

本文编号：1451061