Recognition And Hazard Assessment Of Weibo Rumors Based On Ensemble Learning

With the increase of network information,the supervision and legislation about network rumors have gradually become one of the researches focuses in various fields.Analyzing the definitions and the characteristics of network rumors plays an important role in the recognition of the rumors in network information.And the hazard assessment of network rumors lays a certain foundation for the prevention and control of cyberspace.This thesis takes the Weibo platform as the research object.On the basis of summarizing and analyzing the deficiencies of the existing study on network rumors,the methods of Weibo rumors recognition and hazard assessment are specified.The main work is as follows.Firstly,aiming at the problem that the current definition of network rumors is not clearly defined,the definition of network rumors and Weibo rumors are given in combination with the 5W mode in communication,and the definition of Weibo rumors is formalized.Then,based on the previous research,a feature set suitable for Weibo rumor recognition is constructed and filtered.And a Weibo rumor recognition model based on Stacking Ensemble Learning is proposed.Experimental results show that the Weibo rumor recognition method can effectively identify rumors,and has the best effect in practical applications.Secondly,for the lack of quantitative research on the current evaluation of network rumors,based on the previous study of the index system for public opinion crisis warning,the hazard assessment index system of network rumors is proposed.Using the fuzzy analytic hierarchy process to assign weights of each indicator in the index system,the hazard of Weibo rumors is ranked by applying the Random Forest algorithm.The results show that this method can effectively evaluate the hazard of Weibo rumors.Thirdly,in order to solve the problem that there are no effective means to monitor and control the current mass spread of network rumors,a Weibo public opinion early-warning prototype system,which combines rumor recognition and hazard assessment,is designed and implemented based on the first two aspects of the research work in this thesis.