Research On Identifying Influential Nodes In Complex Networks Based On Network Embedding And Local Properties

Identifying influential propagation nodes in a network is an important research area in complex networks.It involves the structure and function of the network,including degree distribution,connectivity,information dissemination,and robustness.In practical applications,it can control information.Spread on the network,do efficient news promotion,and avoid the spread of faults in the power grid.Most of the existing methods to identify the most influential propagation nodes in the network are based on traditional network representation methods.These traditional representation methods generally have problems such as network sparseness,high dimensions,and high computational complexity.Based on the network embedding method,combined with the local topology and similarity of the nodes,this paper proposes three related algorithms to identify the most influential nodes in the network.The main work and results are as follows:1.A method for identifying influential propagation nodes based on DeepWalk and local centrality is proposed.The DeepWalk network embedding algorithm is used to map the high-dimensional complex network to a low-dimensional vector space,calculate the Euclidean distance between the local node pairs,and then combine the network topology information to calculate the influence of the desired node on all nodes in the neighborhood.the summation is used as an indicator to judge the influence of the node size.finally,using different central methods to calculate the infection ability and kendall coefficient of top-10 nodes and top-10 nodes in real networks,it is shown that the method in this paper performs well in accuracy and stability.2.A method for identifying influential propagation nodes based on the effective propagation of edge is proposed.The method holds that each edge has different effects on the propagation of information because of the different nodes connected by it.Therefore,it is a feasible strategy to solve the problem of maximizing the influence by considering the influence of each edge on the node propagation.the simulation results in real networks show that the algorithm in this paper performs well in accuracy and time complexity.3.A method for identifying influential propagation nodes based on SDNE and clustering algorithm is proposed.In this method,the SDNE network embedding is used to obtain the low-dimensional vector representation of each node in the network.then,the clustering algorithm is used to divide these vectors into k associations.finally,according to the local topology of the network,the leader node of each community is obtained,that is,the Top-k node in the influence maximization problem.The experimental results of using IC model,LT model and number statistics of selected nodes in different data sets show that the method in this paper performs well.