Analysis of the Evolution Trend of Segmentation of Online Public Opinion Based on Dynamic Tracking Technology

Abstract

Abstract In this paper, the dynamic tracking technique is applied to analyze the triggering factors of online public opinion, modify the feature subsets and establish a basic database of sensitive information. The test of statistical characteristics of the evolution process of public opinion is realized by using the prediction of time series by the ANOVA algorithm, and on this basis, the smoothness of the model is realized by testing the parameters. In the process of establishing the quantitative index system, the seasonal difference algorithm is introduced to predict the development trend of public opinion on hot topics to realize the visualization model design. Finally, in order to verify the validity of the model, an experiment was conducted with Event A as an example. The results show that users can set a total of 4 keywords with an average word frequency of 0.03 Hz. The prediction accuracy in different periods can reach 88.73% on average, and the adaptation value can reach 0.8 percentage points on average. Thus, it can be seen that the trend analysis model of online public opinion evolution constructed in this paper can quantitatively analyze the mechanism of public opinion segmentation evolution and the trend prediction problem and provide a theoretical reference for the government to govern public opinion.