Micro-blog Hot Events Prediction System Based On Information Propagation Model

In recent years,with the development of Internet technology,the user groups of social media have gradually grown.Nowadays,online social networks have quickly become very important hubs of social activity and conduits of information.Weibo is one of the most widely used social platforms in China today and the users of weibo are huge.And weibo's hotspot events forecast,especially the prediction of emergencies becomes an important task of public opinion monitoring.However,some current research methods do not consider graph relationships among user nodes in the social network,leading to poor prediction results;And some prediction methods simply consider the graph model relationships among user nodes,and do not take into account the impact between users,also leading to poor prediction results.This article focuses on these issues to carry out microblogging hot event prediction research.On the basis of analyzing and summarizing the existing propagation models and prediction algorithms,this paper designs and implements a better micro-blog hot events prediction system based on information propagation model.And the system's hotspot prediction effect is experimentally tested.The main work of this article is as follows:First,This paper proposes an information dissemination model based on user node influence and IC model.The current information communication model has the defect that users' influence is only calculated based on the number of fans and the user logo,this paper calculates the users' influence based on the weighted directed two sub graph of the user-micro-blog group.We Combine with user attention,information sensitivity,and whether the chain contains other factors to carry out the prediction of information dissemination,more in line with the actual situation.Second,the system uses historical popularity to extract keywords,and through an improved information dissemination model,it can accurately describe the predicted hot events.For the characteristics of new words appearing on Weibo,this system proposes an HP-IBF algorithm to extract keywords based on historical popularity and the rarity of words.The experimental results show that the prediction efficiency of the algorithm of this system has a certain increase under the condition of ensuring a certain amount of data.Third,this system uses the SSH framework based on the J2 EE platform,adopts the object-oriented design method and the MVC design concept,follows international industry standards,J2 EE specifications,Web services,AOP,etc.Meanwhile,it also designs and implements the function of user management,algorithm management and dataset management with good scalability,low coupling,safety and technological advancement.Finally,Experimental comparisons and analyses were conducted on real-world microblogging datasets of different scales with other predictive models to verify the feasibility and forecasting effect of hotspot prediction models,and the system we designed can complete the microblogging hot event prediction.