基于国际贸易网特性的贸易依赖派系过滤算法  

Trade dependence weighted clique percolation algorithm based on characteristics of international trade networks

作　　者：郑军[1,2] 张永庆[1] 黄霞[1,3] 沈爱忠[1] 

机构地区：[1]上海理工大学管理学院,上海200093 [2]宁波职业技术学院,浙江宁波315800 [3]江苏科技大学(张家港),江苏张家港215600

出　　处：《计算机应用研究》2017年第12期3708-3712,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(71571119);江苏省现代教育技术研究课题(48888);江苏高等教育科学研究"十三五"规划课题(16YB064)

摘　　要：目前诸多社团划分算法往往面向一般意义的复杂网络,针对具体问题应用时需要筛选和改进。为了提高国际贸易网络社团划分的准确性与合理性,考虑现实国际贸易主体平等、联系普遍、流量悬殊、市场依赖、贸易组织重叠、社团相对稳定和贸易波动等特性对社团划分的影响,提出一种引入贸易依赖度权重的派系过滤算法。将该算法应用于海上丝绸之路贸易社团划分,通过检验算法结果的社团结构可辨识性、模块度Q值、现实契合度和可解释性,验证了该算法的有效性。In allusion to common complicate networks,current community partition algorithms may have limitations in applying to specific problems. In order to partition international trade communities exactly and reasonably,this paper presented an improved algorithm called dependence weighted clique percolation( DWCP). It took into account the influences of characteristics of real international trade networks,in which equal countries widely connect and depend with each other with quite different and fluctuant trade flows,constitute overlapping organizations and relatively stable trade communities. Then,it was used to find the trade communities on the Maritime Silk Road,and compared with several other common algorithms. The result proves the valid of this algorithm through comprehensively testing community structure identification,modularity degree,degree of consistent with reality and the interpretability.

关 键 词：国际贸易网 社团划分算法 海上丝绸之路 贸易依赖度 社团划分检验标准 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] N941[自动化与计算机技术—计算机科学与技术]