Research And Application Of Influence Maximization In Social Network

With the modern internet technology atmosphere, more and more people use to live with social networks. Compared to receive information passively from the traditional media, people are more likely to communicate in the online community. However, relationship between users of these community is very complex so that it is difficult to trace dissemination of information rapidly. With lagging of network real name system, the online community is often filled with rumors. Therefore, study of how to use the features of social network to guide public opinion has important application value and practical significance to inhibit the spread of rumors quickly and efficiently.The individual of social network compose to a group of common characteristic through interact activates such as information exchanging and sharing. It is a hot issue in the study of social networks that how to min influential nodes with the community structure of network and make use of interaction between them to a most effect dissemination and diffusion, resulting in a positive guide to the network.In this paper, technology of complex network studding has been used to establish a model of information communication. Using the topological structure of community network and analyzing of influence of nodes, guiding of public opinion has been realized through blocking or changing the propagation path of target information.First, this paper summarizes the studding of influence maximization, leading to the characteristic of sub-modular of influence function and an improved greedy hill-climbing algorithm based on this characteristic. This improved algorithm overcome two questions that the high time complexity when calculation of node influence is in the whole network and the lack of accuracy due to the limitation of calculation in a sub-graph composed by some of its neighbor nodes.Then, based on the effect of community structure to the process of information dissemination in the network, some important parameters used to compute influence of nodes such as the average effect of node in network have been established in this paper. A calculation model of influence maximization based on the community structure has been proposed, and verified by experimental analysis.Finally, a public opinion guide strategy based on the calculation model is designed and realized in this paper. This strategy has obtained a good effect when it is applied to the public opinion guidance system.