| The change and development of the information technologies further highlight the characteristics of the network environment,i.e.,the openness,the interactivity,and the dispersivity,while it also creates an ideal space to meet the diverse information needs of the network audience group.Under this background,the initiative,the willingness,and the tendency of social people to express their own opinions and emotions about hot events in cyberspace are also gradually increasing,which further makes the uncertainty,the uncontrollability,and the variability of the current network public opinion being obvious.However,the evolution and development process of the network public opinion can have an important impact on the orientation of social public opinion,the government credibility,and the sense of worth of public groups,which means that if the relevant government departments can effectively sort out and control the overall evolution trend of network public opinion,and depict and divine the evolution characteristics of network public opinion,it can have a correct and positive guiding role for the network audience,and then avoid the occurrence of the bad phenomena,including the network rumors,the network violence,and network abuse,notably,it is of great practical significance to the construction of a clear cyberspace and the creation of a healthy and civilized network environment.Based on this,this paper takes “utilizing the parameters characterize the evolution process and characteristics of the network public opinion” as the goal orientation and takes “effectively guiding and controlling the evolution and development of the network public opinion” as the fundamental purpose.At the same time,this paper introduces various strategic methods into this research for forming the information intelligence,including the machine learning,the deep learning,the intelligent optimization,and the system simulation,which aims to provide some supports to enhance the supervision efficiency of the network public opinion for the relevant government departments.Specifically,this paper has arranged eight chapters to explain and discuss the main research content,and the overall research process follows a logic order,i.e.,“deconstruction before reconstruction,from inside to outside,from point to surface”.In that case,it explains the research origin based on the current status of supervision and monitoring of the network public opinion in the first chapter,introduces the relevant conceptual and theoretical problems in the second chapter,expounds the related issues about the evolution characterization parameters and the parameter orientations in the third chapter,demonstrates the quantitative analysis process of the proposed three parameters from the fourth chapter to the sixth chapter,respectively,recombines these three parameters in the seventh chapter,explores the guidance-control strategies of the network public opinion in the eighth chapter,and summarizes the whole research process in the ninth chapter.The core chapters of this paper are the chapter 3,the chapter 4,the chapter 5,the chapter 6,the chapter 7,and the chapter 8,which are presented as follows:The Chapter 3.the source orientation and analysis principle of the evolution characterization parameters of the network public opinionThe main purpose of this chapter is to explain the source orientation of the evolution characterization parameters of the network public opinion,and to clarify the analysis principles of these characterization parameters.Firstly,this paper clarifies the selection mechanism for the source orientation of the evolution characterization parameters of the network public opinion to determine the three source orientation for constructing the valuable characterization parameters,i.e.,the orientation of the information interaction,the orientation of the public opinion reversal,and the orientation of the derivative public opinion.At the same time,this paper proposes the characterization parameters for describing the evolution of the network public opinion under the determined source orientation,i.e.,the information interaction degree parameter,the reversal tendency degree parameter,and the public opinion derivative degree parameter;Secondly,in order to effectively use the proposed characterization parameters to describe and control the evolution process and evolution characteristics of the network public opinion,this paper defines the three analytical principles,i.e.,the quantitative analysis principle of the three proposed parameters,the integrated analysis principle for fused multiple strategies,and the collaborative application principle of the three proposed parameters.The Chapter 4.the quantitative analysis of the information interaction degree parameterThe main purpose of this chapter is to construct the quantitative model of the information interaction degree parameter,in addition,this chapter utilizes the real case to conduct the empirical analysis.The information interaction degree is a parametric variable under the orientation of the information interaction,which is used to measure the degree of information interaction between different network users in the evolution process of the network public opinion,while it is also used to indicate the discussion degree of a certain hot event.From the perspective of interaction and information behavior,this part takes the real comment data of social network users as the research object.In the process of the quantitative analysis,firstly,this paper proposes the novel text similarity calculation method based on the improved Ochiia coefficient to fully explore the potential relationships between the different extracted word texts;Secondly,this paper confirms the cognition tendency categories by introducing the density peak clustering algorithm;Thirdly,this paper obtains the emotional elements required in this section by referring to the existing methods of calculating emotion values and proposes the basic strategy of variance weighted information entropy based on the aforementioned emotional elements.Finally,this paper constructs the quantitative model of the information interaction degree parameter based on the variance weighted information entropy strategy.The Chapter 5.the quantitative analysis of the reversal tendency degree parameterThe main purpose of this chapter is to construct the quantitative model of the reversal tendency degree parameter,in addition,this chapter utilizes the real case to conduct the empirical analysis.The reverse tendency degree is a parametric variable under the orientation of the public opinion reversal,which is used to measure the occurrence opportunity tendency of the public opinion reversal phenomenon or the reversal degree of the public opinion event.From the perspective of the network user group’s emotional emotions,this part also takes the real comment data of social network users as the research object.In the process of the quantitative analysis,on the one hand,this paper aims to construct an emotion recognition model with better generalization performance and better classification performance based on the idea of optimizing the classification model by using the meta-heuristic intelligent algorithm.Specifically,this paper proposes the novel EGSSA algorithm as the optimizer to steadily improve the generalization performance and classification performance of the target emotion recognition model and finally proposes the novel Ms O-KELM model by integrating the EGSSA algorithm with the KELM,which can be taken as the emotion recognition model to identify and classify the emotional feedback of the network user groups in the process of event evolution;On the other hand,this paper constructs the emotional space of network users to judge the group emotion deviation state which appears in the evolution process of the public opinion event,i.e.,utilizing the calculation results of the attribute concepts of the emotion gathering surface,the emotional vertical line,and the emotional fit rate to measure the opportunity tendency of public opinion reversal in this event,and finally constructs a quantitative model of reversal tendency parameter based on these calculation results.The Chapter 6.the quantitative analysis of the public opinion derivative degree parameterThe main purpose of this chapter is to construct the quantitative model of the public opinion derivative degree parameter,in addition,this chapter utilizes the real case to conduct the empirical analysis.The public opinion derivative degree is a parametric variable under the orientation of the derivative public opinion,which is used to measure the probability of occurring the derivative public opinion.From the perspective of the semantic similarity in the text,this part still takes the real comment data of social network users as the research object.In the process of the quantitative analysis,on the one hand,this paper proposes the two-stage density peak clustering based on multi-strategy optimization(TMs DP)to effectively identify the different theme categories formed by the public opinion events in different stages of the life cycle.At the same time,this paper also uses the silhouette coefficient as the criterion for the number of theme categories to extract the theme categories existing in the user comment data;On the other hand,this paper calculates the semantic similarity between the extracted themes and the public opinion events and aims to clarify the derivative topics present in the identified themes.Finally,this part proposes the quantitative model of the public opinion derivative degree parameter based on the similarity calculation results,the role weight of each identified derivative topic in the theme category of the current life cycle stage,and the introduced Bayesian theory.The Chapter 7.the construction and simulation analysis of the network public opinion guidance-control model based on the three proposed characterization parametersThe main purpose of this chapter is to explain the impact of the proposed characterization parameters on the evolution process of the public opinion.First of all,based on the basic principle of the SIR communication model,the information demand theory,the information communication theory,and the motivation cognition theory,this paper clarifies the classification of network user groups in the evolution process of different information transmission states and the state transformation process between different groups,and sets the three proposed parameters as the parametric variables which can affect the information transmission states between different network user groups;In the next place,this paper constructs the network public opinion guidance-control model by utilizing the system dynamic simulation method of the Any Logic software,where this paper aims to clarify the state transmission rate between different groups by clarifying the differential equation and the partial derivative solution process of the proposed model;Finally,according to the specific network public opinion events and the existing research literature,this paper sets the initial value of the basic variables in the network public opinion guidance-control dynamic simulation model and completes the whole simulation analysis process.Moreover,this paper performs the sensitivity analysis for these three parameters and obtains the valid enlightenment from the simulation results.The Chapter 8.the network public opinion guidance-control strategy based on the three selected orientationsThe main purpose of this chapter is to put forward the network public opinion guidance-control strategies,which aims to empower the improvement of the public opinion supervision efficiency for the relevant government departments.This part is mainly to propose the different network public opinion guidance-control strategies under the perspective of the orientation of the information interaction,under the perspective of the orientation of the public opinion reversal,and under the perspective of the orientation of the derivative public opinion,respectively,which are based on the quantitative analysis results of the three aforementioned parameters and the simulation results of the network public opinion guidance-control model.Specifically,under the perspective of the orientation of the information interaction,this paper proposes two significant strategies which are about the information interaction and the information value,i.e.,dynamically perceiving the information value of the network public opinion carrying and effectively identifying the public opinion information spread by network users;Under the perspective of the orientation of the public opinion reversal,this paper proposes two significant strategies which are about the disposal behavior for the public opinion reversal and the emotions of the network user groups,i.e.,comprehensively improving the existing disposal mechanism of public opinion reversal and quick detecting the emotional status of network users;Under the perspective of the orientation of the derivative public opinion,this paper proposes two significant strategies which are about the formation inducement of the derivative public opinion and the monitoring process of the derivative public opinion,i.e.,deeply optimizing the existing response process of derivative public opinion and real-timely monitoring the potential evolution trend of derivative public opinion.In summary,the main researches in this paper are of great theoretical and practical significance.Specifically,on the theoretical level,this paper not only expands the research degree of the network public opinion,but also enriches the method system and theoretical system of the network public opinion research;on the practical level,this paper can not only provide a clear basis for the public opinion supervision of the relevant government departments,but also provide effective decision-making support for guiding,controlling,and governing the network public opinion. |
PDF Full Text Request