Research On Influence Maximization In Microblog Network Based On Improved Independent Cascade Model

The social network is composed of nodes and edges.The spread of information is by means of the interaction between nodes,which is often the formation of “word of mouth”.Information can quickly reach large-scale diffusion in the network based on the principle of it named as viral marketing.The study of how to make a wide range diffusion of information on social networks to achieve the purpose of network marketing has a certain practical significance.Influence maximization is to find k target nodes in the network as the source of information dissemination,and make these nodes to affect other nodes continuously with a certain information dissemination mechanism to achieve the max diffusion of information ultimately.With the continuous development of Internet technology,online social network plays an increasingly important role as a platform for information dissemination in people's lives,which microblog has developed rapidly.This paper studies the influence maximization in microblog network.The greedy algorithm has a high time complexity,and the heuristic algorithm does not perform well in information diffusion.In spite of some improved algorithms,it still cann't meet the specific needs.How to improve the effectiveness of results and reduce the time consumption of the algorithm according to the characteristics of the network platform itself under the large-scale data set is still worthy of study.By analyzing the microblog data,it is found that there are unequal skewed relationship and different availability of nodes in the network.Based on this,this paper improves the calculation of the influence between nodes and constructs the information diffusion of microblog named as IDSA.The algorithm for influence maximization based on IDSA is proposed later,which improves the time-consuming step in the greedy algorithm.The algorithm is implemented on the data set from the Sina and Tencent microblog,and compared with the traditional algorithm.The results show that it has better performance in time complexity and influence range.