Analysis And Prediction Of Internet Rumor Diffusion Based On Representation Learning

With the development of social networks,the number of social software,such as Weibo and BBS,is increasing.The information exchange and resource sharing between people are also increasing.The new generation of social information dissemination technology promotes rapid information sharing and large-scale information cascading.Online rumors can quickly spread their influence and can even spread faster and more widely than real information because of their concealment,suddenness,and dispersion.Therefore,the analysis of the internal propagation laws of online rumors have great significance in guiding correct public opinion,curbing the spread of rumors,and maintaining social stability.This thesis mainly models and analysis the forward and popularity of rumor topics from two levels of individual user behavior and user group behavior.The contribution of this thesis can be summarized as follows:1.In the aspect of individual user behavior,this study considers the diversity and complexity of the rumor propagation feature space and the advantages of representation learning in the feature extraction of data.Further,this study adopts the corresponding representation learning methods for their content and structure of the rumor and anti-rumor to reduce the spatial feature dimension of the rumor-spreading data and to uniformly and densely express the full-featured information feature representation.Second,this paper introduces an evolutionary game theory,which is combined with the user-influenced rumor and anti-rumor,to reflect the conflict and symbiotic relationship between rumor and anti-rumor.Finally,the whole model is proposed.Time slice and discretize the life cycle of rumor is used to synthesize the full-featured information feature representation of rumor and anti-rumor.2.In the user group behavior,aiming at the rumor topic user behavior characteristics,a rumor topic fever calculation method is proposed,and a prediction model based on representation learning and dynamic exponential smoothing algorithm is further constructed.First,considering the timeliness of rumor topic evolution,the life cycle of the rumor topic is time sliced and discretized to obtain the time series of rumor topic fever.Secondly,considering the multi-message characteristics under the rumor topic,using knowledge representation method to learn from different messages.Combined with the variability in the spread of actual rumors,the dynamic smoothing coefficient is used to optimize.Finally,a rumor topic fever prediction model is constructed,and the rumor topic fever prediction is performed in real time,so as to realize the dynamic description and monitoring of the rumor topic propagation situation.At last,the experimental verification is carried out with the data set of microblogs.The experiments denote that the model can not only effectively analyze user group behavior regarding rumor but also accurately reflect the competition and symbiotic relation between rumor and anti-rumor diffusion.