A Study On Prediction Of Micro Blog Negative Information Forwarding Behavior

With the rise of Web2.0,Micro Blog,as an important platform and form for users to create content,has been developed rapidly,and its characteristics such as low threshold,openness and easy-to-use lead to the incregoogle translateasing spread and exchange of information based on Micro Blog.Due to the different motivations of Micro Blog users for creating,publishing and disseminating information,as well as the different users' information literacy and ability to identify information,a large amount of negative information is generated on Micro Blog and quickly spread after multiple reposts,bring challenges to the management of Micro Blog and even cyberspace.Users' forwarding behavior is the main way to spread negative information through Micro Blog network.Starting from the Micro Blog negative information forwarding mechanism,the factors influencing the Micro Blog negative information forwarding behavior were analyzed in this paper,and the prediction study on the Micro Blog negative information forwarding behavior was carried out from micro and macro levels.Firstly,based on the theory of planned behavior,the influencing factors of Micro Blog negative information forwarding behavior were analyzed from three aspects: attitude,subjective norm and perceptual behavior control,and a model of influencing factors of Micro Blog negative information forwarding behavior was constructed and the influencing factors were mathematically expressed and measured.Secndly,a prediction model for Micro Blog negative information forwarding probability was constructed based on the ibnary logistic regression model to predict a single user's negative information forwarding behavior from the micro level;on this basis,a prediction model for negative information forwarding scale was constructed was constructed to predict multi-user's negative information forwarding behavior from the macro level.Finally,the hypothesis test and simulation test were performed on the influencing factor model and the prediction model respectively.The results of hypothesis test show that the influencing factors of the negative information forwarding behavior proposed in this study are indeed important factors for studying the negative information forwarding behavior.The results of simulation test show that the prediction model for Micro Blog negative information forwarding probability is high in accurate and good in performance;the prediction model for negative information forwarding scale not only has high prediction accuracy,but also can better simulate the spreading rate and range of negative information in real Micro Blog network,and effectively predict the evolution trend of Micro Blog negative information forwarding scale.Due to the protection of user privacy in Micro Blog platform,a small amount of data was used in simulation test in this study.At the same time,it failed to take full account into the timeliness of the spread of negative information in the perdiction of the negative information forwarding scale.Therefore,the prediction model in the future research can introduce time series to make real-time dynamic prediction on the forwarding scale,and try to expand the data volume of the experimental set to further improve the practicality and accuracy of the prediction model.