链路预测下能源供应链网络合作演化机制研究
发布时间:2018-02-02 17:46
本文关键词: 供应链网络 合作演化 链路预测 网络结构 能源供应链 相似性指标 精确度 耦合 出处:《智能系统学报》2017年02期 论文类型:期刊论文
【摘要】:应用供应链网络结构或节点的属性信息预测未产生链接的节点企业合作的可能性是链路预测应用供应链网络合作演化分析的关键,利用链路预测的理论框架和评价方法,借助5种相似性指标对能源供应链网络合作连边演化进行预测。研究结果表明:当使用供应链网络结构属性作为单一相似性指标时,RWR指标的预测效果最好;耦合指标预测精确度要比单独考虑单一指标时有显著提高;RWR指标和Katz指标耦合效果要优于和CN指标、PA指标、LP指标耦合效果,且RWR指标在耦合算法中起到主导性作用;与直接建立网络演化模型相比,链路预测在分析供应链网络合作演化机制上更为有效。
[Abstract]:The application of the network structure of supply chain or the attribute information of nodes to predict the possibility of the cooperation of node enterprises without producing links is the key to the analysis of the evolution of cooperation in the application of supply chain network in link prediction. The theoretical framework and evaluation method of link prediction are used. With the help of five similarity indicators, the evolution of energy supply chain network cooperation continuity is predicted. The results show that: when the supply chain network structure attribute is used as a single similarity index. The prediction effect of RWR index is the best. The prediction accuracy of coupling index is significantly higher than that of single index. The coupling effect of RWR index and Katz index is better than that of CN index, PA index and LP index, and RWR index plays a leading role in the coupling algorithm. Compared with direct network evolution model, link prediction is more effective in analyzing the evolution mechanism of supply chain network cooperation.
【作者单位】: 桂林电子科技大学商学院;
【基金】:国家自然科学基金项目(71662007) 广西哲学社会科学研究课题(15BJY016) 桂林电子科技大学研究生教育创新计划项目(2016YJCX61)
【分类号】:F274;F426.2;TP301.6
【正文快照】: ZHANG Xuelong,WANG Junjin(School of Business,Guilin University of Electronic Technology,Guilin 541004,China)预测是所有的科学学科所不能回避的问题。链路预测是数据挖掘的研究方向之一,尤其在计算机领域早有较深入的研究,其研究思路主要是基于马尔可夫链和机器学习[1,
本文编号:1485095
本文链接:https://www.wllwen.com/guanlilunwen/gongyinglianguanli/1485095.html