Research On Node Importance Ranking And Influence Maximization Based On Feature

With the rapid development of information technology,the influence of Internet permeates into every aspect of life.Measuring the importance of nodes in the network and revealing the characteristics and mechanism of important nodes in the transmission of influence are of great significance to ensure the stable operation of the network and improve economic benefits.Generally,scholars use the topological structure of the network to design node importance ranking algorithm and influence maximization algorithm.Considering that the node attributes,user preferences and other feature information in the network play an important role in node importance ranking and influence dissemination,this paper introduces feature and feature similarity,and design a new node importance ranking algorithm and a new influence maximization algorithm.The main work of this paper is as follows:(1)A new feature-based node importance ranking algorithm is proposed.The traditional Page Rank algorithm only focuses on the network structure,which lead to inaccurate ranking results.By introducing the node attributes and the user preference features,a feature-based Page Rank(FBPR)algorithm is proposed to identify important nodes accurately and efficiently.The weight matrix and fixed jump probability vector in FBPR model are redesigned by feature similarity.For different application scenarios,different ranking results can be obtained by adjusting node attribute factor and user preference factor.Numerical simulation shows the effectiveness of the algorithm.(2)A new feature-based influence maximization algorithm is proposed.The traditional influence maximization algorithm only focuses on the network structure,and the seed nodes selected are not accurate,and the information transmission model is not appropriate,resulting in insufficient influence.Using FBPR algorithm to select seed nodes,a new feature-based IC(FBIC)model is designed to simulate the process of information transmission,and a feature-based IM(FBIM)algorithm is proposed.The influence matrix of nodes is designed by feature similarity,and the opinion dynamic is used to simulate the change of nodes’ opinion to information.Different influence effects can be obtained when individual opinions receive different information at the initial moment.Experimental simulation verifies the accuracy of FBIM algorithm and further verifies the effectiveness of FBIC algorithm in influence maximization.