Identifying Multiple Influential Spreaders Based On Region Density Curve In Complex Networks

With the development of complex network science,it has become an ef-fective tool to study various complex systems.Using the theory and method of complex network to explore the complex system has important theoretical significance and practical application value.Identification of multi-influence n-odes is a one of hot issue in the field of complex network research,that is,finding multiple nodes in the network to maximize the impact on the net-work.Identification of multi-impact nodes is of great significance for disease control,product marketing and promotion,information dissemination and oth-er aspects in real life.At present,some effective identificat,ion methods have been proposed,such as taking the node with the highest centrality index as the multi-influence node,but this method is relatively clustered in selecting nodes and lacks consideration of the interaction between nodes.Therefore,this paper considers from the characteristics of multi-influence nodes,that is,nodes should be scattered and important enough,so as to ensure the non-redundant propagation and the wide propagation range.This paper proposes a method to identify multiple influential nodes in a complex network based on the regional density curve.The main research includes the following two aspects:1.Considering the network in the core-periphery structure or community structure,first of all to find such a structure,in detecting the mesoscale struc-ture of such a unified method,namely through mapping network area density curve,and then according to the characteristics of the curve to find core n-odes or community internal network nodeand then find the periphery or core and a bridge between community and community nodes,finally determine the influence of the selected node.2.The effectiveness of the proposed algorithm was verified on different networks based on disease transmission model and rumor transmission model,and was compared and analyzed with degree,betweeness,k-shell,degree dis-count,coloring and other centrality indexes.The results show that the method based on the region density curve can better identify the multi-influence nodes in the network.