Initial Nodes Selection Algorithm For Behavior Diffusion Within Social Network

To promote public welfare new behavior in social network is conducive to social self-governance, and research on the most influential nodes in behavior diffusion is of both theoretical significance and realistic value. Behavior diffusion has an essential distinction with information and disease diffusion. Behavior diffusion is not simple contact diffusion, but of repeatedly strengthen, therefore the form of behavior is complex and slow. This paper designs an algorithm of initial nodes selection based on the feature of behavior diffusion.First, starting from the typical application scenarios, four technology requirements are proposed in this thesis: propagation characteristics of different types of behavior, behavioral propagation mechanism, building of real weighted network, single and multiple initial nodes selection algorithm. Second, according to the characteristics of behavior diffusion, a behavioral propagation mechanism is designed. And then a weighted social network based on real society — natural growth model is built in this thesis. Third, this thesis proposes single and multiple initial nodes selection algorithm depending on different needs, of which the single initial node selection algorithm combines the node strength and position in network, and the multiple initial nodes selection algorithm is to avoid redundancy while maintaining effectiveness.Finally, this thesis constructs experimental analysis with NetLogo, and compares to other algorithms from the dissemination efficiency and scope to verify the validity of the algorithm. The results show that: in the case of a single initial node, nodes with a high strength level and network core reach the best propagation efficiency and range; but in the case of multiple initial nodes, to prevent waste of resources, avoiding redundancy can increase propagation efficiency and range.