Research On The Influence Maximization Algorithm In The Polarity-related Social Network

In recent years,the search for lost contact with friends in expanding social circle,and the excellent performance and maintain contact between friends,some large social networking sites,such as Facebook,Twitter,was a huge success.Because can make information and technology rapidly affect the network in the crowd,the social network has become a huge spread of information and product marketing platform,in order to explore the social network as the platform of information diffusion potential,there are a lot of problems to solve,the largest effect is one of the important issues.The social network influence maximization problem in viral marketing and messaging are studied,in order to quickly and accurately select nodes in the social network's most influential collection,after making it through influence propagation model of propagation effects in the whole social network benefits maximum.Study on the influence maximization problem based on a specific social network,the relationship between social network user members of the previous studies are mostly based on the always positive influence,after the user node is activated only in a state and the entire social network activation of the node state are the same,we call this kind of social network for unsigned social network.But the relationship between social network nodes in practice has both positive and negative impacts,nodes through positive and negative influence of the activated state is different,we call this social network as a symbol or polar social network.Because of the different structure of the social network and the social network,the original communication model can not be applied in the polar social network,and the corresponding algorithm for maximizing the impact should be improved.Moreover,the impact of the initial collection of the same node is different,which is a problem to be solved.In order to solve the problems and challenges in the study of these effects,this paper studies the following two aspects:(1)For the unsigned social network and polar social network structure,communication model is not suitable for problems in social networks influence polar correlation maximization algorithm,we extended the IC model,adding quality factor q,proposed a new communication model of IC-P,based on this model,we will give the greatest positive impact PIM algorithm and negative influence maximization algorithm NIM algorithm in order to reduce the time overhead in the algorithm is improved by adding sub models of time efficiency of the algorithm,and finally verified by experimental data.(2)Aiming at the problem of accuracy of perception under different themes,although there are a large number of research topic preference,but are unsigned social network based on social network in polar influence maximization algorithm corresponding to the relative lack of,based on the polarity of social network into thematic preference factors,effect of polar social network under the theme of the perception of the maximization problem.The use of customer intimacy and topic similarity calculated transmission probability of different theme,by setting the threshold of node(?)t.The theme of pretreatment,delete the theme with a lower sensitivity node,improve the efficiency of the algorithm.