Research On Network Public Opinion Crisis Early Warning Based On System Security Dimensionality Reduction Theory

The internet public opinion crisis can be regarded as a kind of complicated public security problem of Galway in terms of its triggering subject,content and communication mechanism.Therefore,it is of great theoretical and practical significance to study the method of simplifying the difficulty of early warning of Internet public opinion crisis.Based on the methods of system security dimension reduction and hierarchy analysis,this paper analyzes the multi-dimensional structure of emergent network public opinion events,and constructs the early warning index system of network public opinion crisis and its early warning methods Transforming the high-dimensional network public opinion crisis and its influencing factors into relatively simple and concrete low-dimensional early-warning indicators,and making it easier to scientifically understand and quantitatively interpret the transformed early-warning indicators,make it easier to warn people.The main findings are as follows:(1)From the perspective of system security,the multi-dimensional structure of Internet public opinion is condensed from the three dimensions of subject,content and communication,that is,the main dimension of public opinion,content dimension,communication dimension of the three first-level factors and its extension of the secondary factors.The results show that the network public opinion crisis has a remarkable multidimension,which is affected by several or all factors in the multidimension structure,and the final dimension influence factor also has the multi-dimension generally.For example,the attitude tendency of the audience in the dimension of the subject of public sentiment,different individuals have different attitudes and motivations(that is,they have the randomness),and the category and nature of the subject matter in the dimension of content are also very different,etc.(2)Based on the theory of system security dimension reduction,the early warning index system including 3 first-order dimension indexes and7 second-order dimension indexes was constructed from the multidimension structure of the factors influencing the network public opinion crisis,and carries on the weight analysis to the final grade index.Then,by analyzing the frequency of each index of network public opinion event sample,the evaluation criteria of public opinion sub-dimension earlywarning index grade are calculated,and the single-dimension earlywarning grade judgment of network public opinion is realized.The results show that the risk level of network public opinion events is a complex comprehensive measure,which should be determined according to the lowdimensional index early-warning set.Finally,the expert assignment method is adopted to determine the level of public opinion early warning by calculating the weighted value E of public opinion comprehensive index.The concrete process is: 1)setting the early warning interval,2)calculating the comprehensive index of public opinion early warning weighted value E,3)judging the value e to fall into the early warning interval.(3)The early-warning factor set U,the early-warning evaluation set W,the early-warning weight set A and the whole-dimension early-warning grade evaluation four modules are determined,and the early-warning model of network public opinion crisis based on the theory of reducing dimension of system security is established.The model can transform coarse-grained public opinion information into medium-fine-grained public opinion information through three stages of data collection,processing,analysis and application,then it is transformed into a dimensionality reduction process of fine-grained public opinion information which is easy to quantify and explain.Finally,11 cases of public opinion on the Internet were verified.After analysis,the earlywarning results of the cases were consistent with the characteristics of the early-warning level of public opinion stipulated by the administration department of Public Opinion on the Internet in China,the practicability and feasibility of the early warning method are verified.Figure37,table12,reference109...