Research And Development Of Social Network Influence Analysis System

The social network has developed along with the development of the Internet,and it has been integrated into people's lives.Our research and understanding of social networks have also rapidly deepened.The problem of maximizing influence is about the spread of information on social networks.We study this problem to make theinformation dissemination tothe widest scope,and the key to this problem is how to choose the user who initially disseminated the information.Among them,influence maximization(IM)is a major aspect of social network influence research.The problem of influence maximizationinvolves finding a certain number of seed users in social networks.These users have a higher influence than other users,and through these users,influence can be most widely spread on social networks.In social network,theremay be some negative influence,it is necessary to spread positive influence to block the negative one in the network.This is the problem of negative influence blocking.These two probelms are important issues in the field of socialnetwork analysis,.Although they are highly theoretical and complex,it can be applied in many fields-,such as online marketing,users recommendation,public opinion analysis,rumor suppression,etc.This articlestudies the spreading of influence in social networks in two aspects:influence maximization and negative influence blocking maximization.Our purpose is to improve the competitiveness and economic benefits of businessmen or companies in viral marketing in social networks.We have alsodeveloped a corresponding simple marketing system suitable for practical applications.The main contributions of our work areas follows.(1)In order to improve the maximum economic benefits of businessmen when conducting viral marketing strategies,an influence maximization algorithm based on experimental design is proposed.The basic steps are as follows:First,treat each user in the social network as an experiment in the classic experimental design problem,and transformthe IMproblem of selecting kseed nodes from the social network to the experimental design problemof selecting the most representative kexperiments.Thenwe design the corresponding impact matrix to represent the entire social network,and construct a mathematical model and optimize it.To solve the optimization problem,we use the method called cross-iteration to get the final parameters.We prove the feasibility of our algorithm and show that the method can achieve better results than other methods(2)To improve the merchant's ability to target competitors' marketing strategies and formulate their own marketing strategies,a method for blockingthe negative effects underthe unified communication model is proposed.The main idea of this method is to extend the unified communication model to competitive social networks with positive and negative effects.By studying the state changes of each node during the propagation of each hop in the network,the propagation model is defined and the correlation is set in the competitive network.This method uses the variables of relationships to quantify the spread of influence.Finally,we use greedy approachto solve the problem.(3)Based on the perspective of practical applications,a network marketing systemwas designed and developed based on network analysis.The system uses an experimental design-based impact maximization algorithm and a unified propagation model-based negative impact blockingmethod.Through the requirements analysis,we propose the overall design of the system frameworkand the module structure of the system and introduce the details ofeach module.