基于深度学习的云模式机场客户行为预测  

Cloud based airport customer behavior prediction and personalized service design based on deep learning

作　　者：段冬生 Duan Dongsheng(Engineering Construction Headquarters of Guangdong Airport Management Group Co.,Ltd.Guangzhou,Guangdong 510470)

机构地区：[1]广东省机场管理集团有限公司工程建设指挥部,广东广州510470

出　　处：《现代科学仪器》2024年第5期315-320,共6页Modern Scientific Instruments

摘　　要：为提高航空运营管理效率,研究从云模式机场客户行为分析的角度,选择改进支持向量回归模型,预测机场旅客出行指数,选择先来先服务算法进行座位的推荐与分配。结果显示,当航班数为525次时,先来先服务算法的整体平均满意度大于0.800,大于历史座位分配结果。云模式下的管理效率得分为0.75,比非云模式下大0.22。云模式下的满意度比非云模式下大0.24。研究方法能有效预测旅客出行指数,可促进旅客整体平均满意度的提高。In order to improve the efficiency of aviation operation management,research is conducted from the perspective of cloud based airport customer behavior analysis.An improved support vector regression model is selected to predict the airport passenger travel index,and a first come,first served algorithm is chosen for seat recommendation and allocation.The results show that when the number of flights is 525,the overall average satisfaction of the first come first served algorithm is greater than 0.800,which is higher than the historical seat allocation results.The management efficiency score in cloud mode is 0.75,which is 0.22 higher than that in non cloud mode.The satisfaction level in cloud mode is 0.24 higher than that in non cloud mode.The research method can effectively predict the passenger travel index and promote the overall average satisfaction of passengers.

关 键 词：云模式 深度学习 客户行为 出行指数 

分 类 号：TP315.69[自动化与计算机技术—计算机软件与理论]