Public Opinion Analysis Application Based On Two Types Of Statistical Models

In the Internet age,social information on the Internet has exploded.The Internet has become an information distribution center and a platform for the propagation and spread of rumors.The complexity of network information has caused many urgent problems in the field of public opinion analysis.This paper uses an adaptive enhancement algorithm to improve the performance of hidden Markov model text classification.Different from the previous sentiment classification models based on traditional statistical models,the hidden Markov model can explore the evolutionary laws of different sentiment texts,combine statistical models with text semantics,and perform sentiment classification with better classification accuracy.This model can overcome the cumbersome and lagging problems in the construction of emotional dictionaries,and can reflect text emotions in a timely manner through the implicit text evolution law.Experiments show that this model has higher classification accuracy than traditional sentiment classification models,and has comparable performance on the test set and training set,and has better applicability.Secondly,this paper proposes a new parameter estimation method for Logistic model.The spread of rumors conforms to Logistic growth.In order to better explore the variables of the model,this paper improves the estimation method of Logistic model parameters to improve the accuracy of the model.The parameters are estimated by variational inference and verified.It is found that compared with the traditional parameter estimation method,the variational inference method only needs a few iterations to obtain the high-precision parameter estimation of the Logistic model.This method greatly improves the estimation efficiency.Finally,due to the rapid growth of Internet information and the large capacity,the scope of the spread of this event has changed during the spread of Internet rumors.This phenomenon does not meet the basic assumptions of the classic communication model.Therefore,this article improves the upper limit of the growth of the SIR model..Then the factors that affect the spread of rumors are divided into exogenous variables and endogenous variables.The former is controlled by the media’s response speed to rumors and media influence,and the latter is determined by the mutual influence of groups.On this basis,the relevant parameters of media influence on the spread of rumors are analyzed through models.Media with high influence need to pursue response speed when facing rumors,while media with small influence need to pay more attention to the authenticity and reliability of reports.