Diffusion Tracing And Path Construction For Public Opinion Management

Public opinion control in social media platforms is an important branch in the field of information security,which involves information diffusion tracing,diffusion path construction,topic prediction and hot spot mining.This paper will focus on two aspects of work which includes information tracing and propagation path construction.1)In the part of the research on information tracing,this study designs a user interest-based information tracing algorithm.The algorithm utilizes user interest to calculate the influence of bloggers,commentators,and forwarders.Then,combining user attraction,time parameters and influence mentioned above,rank algorithm will be employed to calculate the score of information and the information source will be found by the score rank.Compared with the existing study,this algorithm improved the recall rate by about 19%.2)In the research of information diffusion path construction,this study proposes an individual attributes-based information path construction algorithm and a user community attributes-based information path construction algorithm.Individual attributes-based algorithm uses interest similarity,topic similarity,user interaction and text similarity among users to calculate the information diffusion probability.Finally,the entire information diffusion path can be obtained by calculating the diffusion probability among all pair users.Experimental results illustrate that the algorithm promotes the path topic similarity,path user number and running efficiency.User community attributes-based algorithm introduces the recommendation system into the construction of information diffusion relationship and simulates the process of information diffusion between users as the process of information recommendation between users,so as to judge the user information diffusion relationship by recommendation system and acquire the whole information diffusion path.Experiments based on two types of data sets implies that the results of the algorithm are optimized in accuracy,precision,recall rate and F1 value,and its path structure satisfies the relevant properties of information diffusion.