Research On Information Diffusion And Influence Maximization In Social Network

At present, to study the influence maximization in the social network mostly through selecting the most influential k nodes in the network as the initial node activation. However, for a specific social network users, the network topology is not open to him, what he only knows is limited to its direct link neighbor users, so when the user wants to spread a piece of their own information, a more reasonable model is to select a finite number of neighbor nodes to help them, rather than through selecting k nodes in the whole network. Based on the traditional network global influence maximization algorithm about how to choose the most influential nodes in the network have been fruitful, so we consider the design of an algorithm, which makes the neighbor nodes selection means can relate to the traditional algorithm of the selected nodes. By indirect activation of these influential nodes, the user's influence would get the maximum diffusion in the network.The main innovations of this paper are as follows:1.presenting closeness dependency degree to assess the user's neighbor's location to the network's most influential nodes, and combine the index with the traditional greedy algorithm. We call it PIMCD algorithm, we set to practice in four different characteristics of the data for the experiment, the experimental results show that the PIMCD algorithm perform better in the influence scope and time than the greedy algorithm and the existing heuristic strategy, especially in the structure is relatively dense networks.2.in view of the PIMCD algorithm in the structure is relatively sparse network of underperforming shortcomings, we designed the new evaluation index which was used to evaluate the effects of node diffusing capacity,a new algorithm PIMCAC. we also conducted experiments on four data sets, it is proved that the algorithm can effectively solve the user personalized structure in relatively sparse network.In this paper the influence of user individuation has made the preliminary discussion on the information diffusion field, establish the foundation for follow-up study of the same field.