Research On The Construction Of Topic Graph Of Social Network Public Opinion Events And Ecological Governance Of Public Opinion

Since the 18th and 19th National Congresses of the Party,the Party Central Committee has attached great importance to the governance of online public opinion dissemination and the construction of public opinion ecology,and has repeatedly put forward important instructions and requirements for strengthening online social management and scientific guidance and governance of online public opinion in various conferences.Social network opinion events are dynamic microcosms and concentrated manifestations of social events,and they also have a certain influence on people’s opinions and the direction of public opinion and ideology in the real society.On the one hand,public opinion events can highlight public attitudes and expectations and promote the handling and resolution of events.On the other hand,the accumulated negative and irrational information in them may intensify conflicts in social events and jeopardize social stability and harmony if they are not controlled timely and effectively.Knowledge graph provides an efficient technical means for online public opinion research and facilitates the application and analysis of unstructured information.With its advanced knowledge organization,rich semantic information expression and open interconnection characteristics,topic graph of public opinion events can express and analyze events from different elemental dimensions.It helps to obtain comprehensive insight and understanding of public opinion events in a scientific and effective manner and propose dynamic and specific ecological governance strategies for public opinion.Based on this,this paper takes social network public opinion events as the research object,adopts literature research,empirical research,knowledge graph and machine learning to construct topic graph of different dimensions of events based on the perspective of information ecology theory,constructs an ecological governance model of public opinion based on this and proposes ecological governance strategies for public opinion.Based on the introduction in Chapter 1,which introduces the background and significance of this paper,sorts out the current status of domestic and international research,clarifies the research content and methodology,and introduces the relevant concepts and theories in Chapter 2,the subsequent six core chapters are carried out.Chapter 3 is the theoretical core of this paper,which proposes the analysis model of social network opinion event topic graph.Based on the analysis of event elements and relationships,the event elements are corresponded to information ecological elements.It also points out that the topic graph of social network opinion events includes three thematic dimensions:user population graph,information community graph and spatio-temporal evolution graph.Secondly,the "Japanese nuclear waste water discharge" incident is taken as a typical case.Chapters 4,5 and 6 use information person,information and information environment as the basis to construct and analyze user population graph,information community graph and spatio-temporal evolution graph of public opinion events according to the logic from microscopic to mesoscopic to macroscopic.The user population relationship and structure,information focus and topic evolution,spatio-temporal layout and evolution pattern of public opinion events are clarified to provide theoretical support for chapters 7 and 8.Then,Chapter 7 constructs the ecological governance model of public opinion based on the results of graph analysis from Chapters 4 to 6,and provides theoretical support for the proposed ecological governance strategy of public opinion in Chapter 8.Finally,Chapter 8 proposes the framework of ecological governance of social network public opinion and specific governance strategies on the basis of Chapters 4 to 7,which is the landing point of this paper at the practical level.the details of the six core chapters are as follows:Chapter 3 Construction of Topic Graph Analysis Model for Social Network Public Opinion EventsThe model of elements and relationships of social network public opinion events is proposed.The event is divided into four elements:subject,ontology,space-time and carrier,and the elements and relationships of public opinion events are analyzed to clarify the structure of public opinion events.Secondly,based on information ecology theory,information person,information,information environment and information technology are corresponding to the four elements of events,and a topic graph analysis model of social network opinion events is constructed.It is pointed out that the event topic graph is composed of three dimensions:user population graph,information community graph and spatio-temporal evolution graph.Finally,a governance model of social network opinion ecology is constructed based on the analysis results of the topic graph to provide a basis for the formulation of governance strategies.Chapter 3 is the core theoretical framework of the whole text,which elaborates the construction and analysis logic of the topic graph of social network public opinion events and provides the theoretical main line for the research work in the subsequent chapters.Chapter 4 Construction of User Group Graphs of Social Network Opinion EventsThis chapter analyzes public opinion event users based on the information person factor.The attribute features of users involved in public opinion events are divided into personal attribute features including number of tweets,number of followings,number of followers and microblog rank.And behavioral attribute features,including sentiment and influence,and quantified behavioral attribute features by Convolutional Neural Network and weighted PageRank,respectively.After standardizing the quantitative data of user attribute features using the Z-score normalization method,user groups are obtained by the Canopy+K-means clustering algorithm,and the user group composition,relationships,sentiment structure and relationship structure of the overall event and each sub-event are visualized and analyzed by graph.This chapter analyzes the subject element of users in public opinion events at the micro level,and lays the foundation for analyzing information at the meso level in Chapter 5.It also provides theoretical support for the analysis of the ecological governance model of social network public opinion in Chapter 7 and the ecological governance strategy of public opinion in Chapter 8.Chapter 5 Construction of Information Community Graph of Social Network Public Opinion EventsThis chapter analyzes public opinion information based on information factors.Extracting keywords and constructing an evolutionary graph of information features using TF-IDF to clarify the focus of information in public opinion events and its dynamic changes.Then the topic communities are obtained through the construction of co-word networks,network analysis and Louvain community discovery algorithm,and topic community graph is constructed to clarify the association relationships between key information features.Finally,the topic content of the topic community is obtained by manual coding,and the topic evolution map is constructed based on the relationship between the keywords contained in the topic community to clarify the topic content and association relationship of the topic community.This chapter analyzes the ontological element of information in public opinion events from the meso level,and lays the foundation for analyzing the information environment of events from the macro level in Chapter 6.It also provides theoretical support for the analysis of the ecological governance model of social network public opinion in Chapter 7 and the ecological governance strategy of public opinion in Chapter 8.Chapter 6 Construction of a Spatio-temporal Evolutionary Graph of Social Network Public Opinion EventsThis chapter analyzes the spatio-temporal element of public opinion based on the information environment factor.The global Moran’s I was obtained by global autocorrelation analysis,which determined the overall pattern of spatio-temporal aggregation distribution of events.Local Moran’s I was then obtained using local autocorrelation analysis to clarify the spatiotemporal aggregation types of events,including four types of high-high(H-H),low-low(L-L),low-high(L-H),and insignificant aggregation.Based on this,the spatial hot spots,cold spots and anomalous regions of the events and their patterns over time were identified.The spatio-temporal network graph is constructed to supplement and deepen the results of spatial autocorrelation analysis,and the local clustering coefficients and flat clustering coefficients of the spatio-temporal network mapping are analyzed with the help of graph algorithms to clarify the network structure and key cities of the spatio-temporal network of events.This chapter analyzes the temporal and spatial elements of public opinion events at the macro level,and provides theoretical support for the analysis of the ecological governance model of social network opinion in Chapter 7 and the ecological governance strategy of public opinion in Chapter 8.Chapter 7 Construction of Ecological Governance Model of Social Network Public OpinionThe construction and analysis of the ecological governance model of public opinion is based on the perspective of the holistic nature of the information ecosystem.The public opinion situation is taken as the target element of the governance model,and 13 secondary element indicators of the model are proposed from the three basic element dimensions of public opinion users(information person),public opinion information(information)and public opinion space-time(information environment),respectively.Combining the statistics and calculations of the topic graph data in the preceding chapters,the Entropy Method was used to determine the weights of each index element,and the public opinion posture curves of the public opinion events were drawn.This chapter echoes chapters 4 to 6 and provides theoretical support for the analysis of public opinion ecological governance strategies in chapter 8.Chapter 8 Ecological Governance Strategies of Social Network Public OpinionThis chapter sorts out the overview,governance objectives and governance framework of the ecological governance of social network opinion.And based on the research contents and conclusions in Chapters 4,5,6,and 7,a framework for ecological governance of social network opinion is constructed.Finally,corresponding ecological governance strategies for social network opinion are proposed from the perspectives of network opinion carriers,users,information,spacetime,and external environment.This chapter is the landing of this paper at the practical level.At the theoretical level,this paper makes a scientific division of event elements for social network public opinion events and proposes a systematic model for the construction and analysis of topic mapping of public opinion events,which provides theoretical support for the construction and analysis of topic mapping of social network public opinion events.Public opinion events are deconstructed and analyzed from three thematic dimensions:user groups,information communities,and spatio-temporal evolution.And on this basis,the results and conclusions of the graph analysis are fused and applied,and the ecological governance model of social network opinion is constructed from the perspective of the wholeness of the information ecosystem.The application of knowledge graph and information ecology theory in the field of social network public opinion research is deepened.Meanwhile,the proposed theoretical model and empirical analysis results lay the theoretical foundation for the development of public opinion ecological governance strategies from different perspectives.At the practical level,the empirical analysis is combined with popular public opinion events to provide practical references for the positioning of public opinion user groups,the detection and evolution of public opinion topics,and the determination of the spatial and temporal evolution patterns and laws of public opinion.The formulation and proposal of governance strategies for public opinion carriers,public opinion users,public opinion information as well as public opinion space-time and external environment in the ecological governance provide methodological guidelines and practical references for public opinion supervisory departments to carry out ecological governance of public opinion.