The Research Of Influence Maximization Algorithm Based On Microblog Network

In recent years,social network plays an increasingly important role in people's real life.How to identify influential user principals in social network and master the evolution trend of the public opinion as well as the guide of conveying information through them,has become the urgent problems at present stage.The research on the maximization influence in social network has always been the focus of scientific research.Although it has achieved very fruitful results,there are still some shortcomings:First of all,most researchers do not consider the potential community structure in the network so that they cannot guarantee the influenced range of seed nodes,secondly,they ignore the influence of transmission characteristics of nodes so that they can not guarantee the influenced quality of seed nodes.Therefore,this paper carried out the following research:1.Airming at the shortcoming of the existing influence maximization algorithm,this paper proposed an algorithm based on maximum influence overlapping community to find the most influencial seed nodes of k.Firstly,this algorithm will find the most influential overlapping communities of k from overlapping communities in the network.Secondly,it will identify the most influential node by measuring and comparing the influence of the nodes from the most influencial overlapping community.And then,the most influencial nodes of k will be found.Compared with the traditional influence maximization algorithm,experimental results show that IMMC algorithm can more accurately model influence communication process and is superior to the competitor in the scope our influence spread.2.Aiming at the limitations of IMMC algorithm,When IMMC measures the impact of nodes in the most influential overlapping communities through community modularity,although it takes the connection between nodes in the community into consideration,it ignores the similarity between nodes,resulting in the imprecise calculation of node's influence.Then,this paper improved existing influence metrics in IMMC and introduced the correction factor based on node similarity which modifies the calculation of community modularity,proposing IMNS algorithm which selected the most influential seed nodes of k more accurately in the network.Experimental results show that the proposed algorithm has better influence on propagation characteristics.