基于支持向量机的“一带一路”航空网络的链路预测  

Link Prediction of"the Belt and Road"Aviation Network Based on Support Vector Machine

作　　者：侯泽林 姚红光[1] 戚莹 HOU Zelin;YAO Hongguang;QI Ying(School of Air Transport,Shanghai University of Engineering Science,Shanghai 201620,China)

机构地区：[1]上海工程技术大学航空运输学院,上海201620

出　　处：《物流科技》2023年第9期80-84,共5页Logistics Sci-Tech

摘　　要：链路预测可以探寻“一带一路”航空网络中的链接偏好,为“一带一路”航空网络提供一定的理论依据。文章选取8个网络结构相似性指标,基于支持向量机分类算法从国家和机场两个层面对“一带一路”沿线航空网络进行链路预测。结果发现:越宏观的航空网络,其预测精确度越高;同时,链路预测模型的预测精确度与指标中包含的网络结构信息相关,指标中包含的网络结构信息越多,预测精确度越高,且对于“一带一路”航空网络中的节点,链路预测中Katz指标的预测准确度最高。Link prediction can explore the link preference in"the belt and road"aviation network,and provide a theoretical basis for"the belt and road"aviation network.In this paper,eight network structure similarity indexes are selected to predict the link of aviation network along"the belt and road"from the national and airport levels based on support vector machine classification algorithm.The results show that the larger the aviation network,the higher the prediction accuracy;at the same time,the prediction accuracy of the link prediction model is related to the network structure information contained in the index.The more the network structure information contained in the index,the higher the prediction accuracy.For the nodes in"the belt and road"aviation network,the Katz index has the highest prediction accuracy.

关 键 词：一带一路 航空网络 链路预测 支持向量机 

分 类 号：F560[经济管理—产业经济]