摘要
针对现有复杂网络节点重要性评价存在的问题:(1)单一指标评价的片面性;(2)多指标评价时各指标选取的主观性;(3)往往忽视指标间的关联信息,该文提出一种基于灰色关联分析的综合评价方法。在对多类常用评价指标的特性分析的基础上,采用极大不相关法实现指标筛选,利用改进熵权法确定各指标的权重,最终以灰色关联分析方法确定各节点的重要性。采用美国航空网数据集进行的实验结果表明,该方法具有较好的节点重要性区分能力,这项工作对于进一步完善节点重要性评价理论框架具有启发意义和实用价值。
To avoid the inadequacy of node importance evaluation in complex networks, such as （ 1 ） the unilateralism of single index ; （2） the subjectivity during index selection of multi-index evaluation （3） the neglect of the correlation analysis on evaluating indexes, this paper proposes a synthesis evaluation approach using the grey relational analysis mechanism. Based on analyzing and comparing the features of several characteristic indexes, infinite irrelevance is introduced to screen indexes. An improved entropy weight calculating method is used to confirm the weight for evaluating indexes and the grey relational analysis is applied to calculate the importance of nodes. A detailed case study on the US air transportation network is given. The experimental result shows that the proposed approach performs better on discrimination and precision. The approach helps to perfect the evaluation of node importance in complex networks.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2012年第4期579-586,共8页
Journal of Nanjing University of Science and Technology
基金
国家自然科学基金(61003210)
江苏省自然科学基金(BK2010491)
江苏省'六大人才高峰'项目(11-C-028)
关键词
复杂网络
节点重要性
极大不相关
熵权
灰色关联分析
complex networks
node importance
infinite irrelevance
weighted entropy
greyrelational analysis
