Research On Method Of Predicting Attention On Health-related Topic Based On Social Network

Social network is a new way of communication that is being popularized rapidly with the development of mobile Internet technology.As one of the most popular social network forms,microblog has gone far beyond the other forms of social networking in many ways,such as the number of users,user activity and so on.With the rapid development of microblog,it has become a source of information with sufficient users.Because of people paying more attention to the changing of health situation and the changing on health attitudes,more and more people publish and exchange information on health-related topics and participate in the discussion of health-related topics on the microblog platform.Therefore,microblog service platform has become an important and rich source of health information and data.Researching on this information and data relating to people's health research topics will be significant meaningful to maintaining people's own health and medical workers'disease prevention,treatment work.Therefore,this paper studies the method of predicting the attention on health-related topic based on the microblog platform.According to the research,this paper chooses Sina Weibo as the source of social network information,and discusses the method to get microblog data from Sina Weibo platform after analyzing its characters.The technologies for predicting attention on health-related topic are also discussed in this paper.Based on studying the features of microblog texts,technologies for tendency of topic attention and topic detection and tracking,this paper establishes and realizes a model for health-related topic detection and attention trend prediction aimed at data from Sina Weibo.The system gets 170828 microblog texts and related data as analytic targets from Sina Weibo by combining the web crawler and Sina Weibo open interface API.After text preprocessing,the system chooses health-related data by SVM classification to guarantee the content of topics be related with health.Then the data is put into the Single-Pass algorithm to detect the health-related topics.This paper defines the calculating method for health-related topic.Then the Markov prediction model is used to forecast attention trend of health-related topic based on the results of topic detection.