Research on the Evolution of Network Public Opinion Based on the Logistic Regression Model

Abstract

The law of network public opinion dissemination is similar to that of infectious disease model. After receiving public opinion information, the netizens are divided into S-type non-disseminator and I-type disseminator. By analyzing the three states of nodes, which are susceptible to infection, transmitted and forgotten, it is used to reflect the communication rules of network public opinions. establish the differential equation of public opinion information spread about time t, according to the general solution formula of Bernoulli differential equation, the prediction function of public opinion dissemination is obtained. The results show that the law of public opinion dissemination follows the Logistics nonlinear regression model, it is a piecewise function about time t. in the rising stage of public opinion dissemination, it follows the SI model, in the declining stage of public opinion dissemination, it follows the improve SIR model. This model can be used to predict the transmission cycle and the peak value of network public opinions, in addition, it can predict the values of relevant parameters : λ、μ、m、n that have decisive influence on public opinion spread. It provides ideas for network public opinion early warning research.