供应商网络结构特征多维层次聚集算法  

Multi-dimensional hierarchical aggregation algorithm for structural features of supplier network

作　　者：萧展辉 唐良运 孙刚 XIAO Zhan-hui;TANG Liang-yun;SUN Gang(School of Computer Science and Engineering,South China University of Technology,Guangzhou 510006,China;Data Platform Business Department of Platform Security Branch(China Southern Network Big Data Center),Digital Grid Research Institute of China Southern Power Grid,Guangzhou 510663,China)

机构地区：[1]华南理工大学计算机科学与工程学院,广州510006 [2]南方电网数字电网研究院有限公司平台安全分公司数据平台事业部(南网大数据中心),广州510663

出　　处：《沈阳工业大学学报》2022年第4期415-419,共5页Journal of Shenyang University of Technology

基　　金：广东省自然科学基金项目(201835426589).

摘　　要：针对传统算法在进行供应商网络结构特征分析时出现的聚集效果较差的问题,提出了供应商网络结构特征多维层次聚集算法.利用Spark框架对供应商网络中部分事实编码进行缓存处理;通过联机分析对编码结果进行转换,采用多维层次聚集算法实现对供应商网络结构特征的多维层次聚集处理.通过对比实验结果证明,该算法能够有效提高聚集效果,且其对供应商网络中无效节点的判断能力较好,能够有效实现对供应商网络结构特征的聚集处理.Aiming at the problem of poor aggregation effect of traditional algorithms in the analysis for the structural features of supplier network,a multi-dimensional hierarchical aggregation algorithm for the structural features of supplier network was proposed.The Spark framework was used to cache some fact codes in the supplier network.The coding results were transformed through online analytical processing,and the multi-dimensional hierarchical aggregation algorithm was used to realize the multi-dimensional hierarchical aggregating process for the structural features of supplier network.The results of comparative experiments show that the as-proposed algorithm can effectively improve the aggregation effect,has better ability to judge the invalid nodes in the supplier network,and can effectively realize the aggregating process of the structural features of supplier network.

关 键 词：供应商 网络结构特征 多维层次 聚集算法 联机分析处理 Spark框架 无效节点 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]