基于数据场的复杂网络节点影响力建模与仿真  

Modeling and Simulation On Influence of Complex Network Nodes Based on Data Field in

作　　者：邵晨曦[1] 陈小齐 王行甫[1] 苗付友[1] Shao Chenxi;Chen Xiaoqi;Wang Xingfu;Miao Fuyou(School of Computer Science and Technology of University of Science and Technology of China,Hefei 230022,China)

机构地区：[1]中国科学技术大学计算机科学与技术学院,安徽合肥230022

出　　处：《系统仿真学报》2020年第7期1257-1266,共10页Journal of System Simulation

基　　金：国家自然科学基金(61472381,61472382)。

摘　　要：复杂网络节点影响力的研究是数据挖掘的重要组成部分。挖掘出复杂网络中有影响力的节点不仅具有重要的学术意义,且有助于抑制流行病的爆发、控制谣言的传播和推广电子商务产品等。通过选取每个节点的混合度分解值(Mixed Degree Decomposition,MDD)作为质量,将复杂网络抽象为数据场,结合数据场模型来识别有影响力的节点,并与一些著名的节点中心性方法进行对比。使用经典的传染病模型(Susceptible-Infected-Recovered,SIR)通过对比感染节点的数量来评估仿真性能。对实际网络的仿真实验结果表明,数据场模型能够有效的识别网络中有影响力的节点。Research on the influence of complex network nodes is an important part of data mining. Mining the influential nodes in complex networks not only has important academic significance, but also helps to suppress the outbreak of epidemics, control the spread of rumors, and promote e-commercial products and so on. By selecting the Mixed Degree Decomposition(MDD) value of each node as its mass, the complex network is abstracted into a data field, the influential nodes are identified by combining the data field model, and some well-known centralities are compares with. The classical Susceptible-Infected-Recovered(SIR) epidemic model is used to evaluate the simulation performance by comparing the number of infected nodes. Simulations on real networks show that the data field can effectively identify the influential nodes.

关 键 词：复杂网络 节点影响力 混合度分解 数据场 仿真 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]