| With the continuous development of network technology,social media has also been enriched.Social media has increasingly become a popular information dissemination platform.Various public opinion events are emerging in an endless stream on the Internet.Due to the anonymity and decentralization of the Internet,netizens can public opinion parties unscrupulously express their thoughts and words,which leads to frequent occurrences of online violence.These cyber violence incidents evolved from online public opinion have spread rapidly,covered a wide range of areas,and suffered great harm to the parties.Cyber violence has once again become a hot topic,and the voice of rejecting cyber violence has become stronger.my country officially promulgated the "Regulations on the Ecological Governance of Online Information Content" on March 1,2020,which explicitly prohibits online violence and human flesh searches,and academic research on online violence has gradually increased.First of all,this article uses relevant theories in the field of public opinion as the theoretical basis to explore the inherent evolutionary laws of cyber violence incidents based on the evolution of cyber public opinion events,and concludes that cyber violence incidents mainly include cyber violence,information exchange and interaction,information diversity,and information.Features such as low retention,vague and volatile event evolution,and rapidity and extensiveness of transmission;from the analysis of evolutionary platforms,it is concluded that cyber violence incidents are mainly based on the Weibo platform as the initial outbreak point,and the transmission speed and efficiency of the mobile terminal are far Higher than the client;from the perspective of evolutionary stages,network violence incidents are divided into public opinion fermentation period,public opinion formation period,public opinion rise period,public opinion transformation period,network violence formation period,public opinion subsidence period,and network violence calming period;from evolution According to the analysis of factors,there are a total of netizens’ attention factors,public opinion reversal factors,public opinion rumor factors,public opinion type factors,and netizens’ value tendency factors.Secondly,this dissertation sorts out the typical cyber violence incidents from 2002 to the present,extracts the relevant characteristics of cyber violence incidents based on the public opinion ontology,public opinion dissemination,and public opinion response,and builds a cyber violence event prediction model oriented to online public opinion.When dealing with cyber violence incidents and general online public opinion incidents,due to the large disparity between the two in the real situation,the imbalance of samples in the model will lead to the weak generalization ability of the model.Therefore,when facing imbalanced data subsets,this dissertation proposes a fusion integrated noise recognition and SMOTE algorithm,which is an improvement of the classic SMOTE algorithm.The experimental results show that the fusion of integrated noise recognition and SMOTE algorithm generates new sample quality is better.Finally,a network violence prediction model is constructed based on a multi-layer perceptron.The accuracy of the model on the training set is 88.7%,and the accuracy on the test set is87.1%,which has good generalization ability.Finally,we will propose targeted prevention,control and governance strategies for cyber-violent public opinion incidents from multiple dimensions.First,based on the analysis of the propagation and evolution of Internet violence public opinion events,provide targeted public opinion management and control strategies based on the evolution characteristics at different stages of evolution;second,based on the proposed cyber violence event prediction model,based on the previous analysis Characteristic importance rankings provide targeted prevention and control opinions from different perspectives. |
PDF Full Text Request