基于聚类与SVR的地区支线航空客运市场需求预测  

Prediction of regional aviation market demand in specific region based on clustering and SVR

作　　者：徐梦瑶 赵鸣 李洋 安洋 张友浩 XU Mengyao;ZHAO Ming;LI Yang;AN Yang;ZHANG Youhao(School of Air Transportion,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

出　　处：《智能计算机与应用》2019年第5期231-235,共5页Intelligent Computer and Applications

基　　金：国家社科基金项目(15BJL104)

摘　　要：针对支线航空客运市场需求预测问题,某些地区(如海南)缺少足够的历史数据,难以建立准确的预测模型。本文提出基于聚类与支持向量机回归(Support Vector Regression,SVR)预测此类地区航空客运市场需求的方法。首先,基于中国各个地区支线航空客运市场需求的分布比,找出与海南分布比相似的地区,再应用系统聚类法在这些地区中找出与海南聚为一类的地区,作为类比地区。然后,选择类比地区的数据样本,通过K-fold交叉验证(K-fold Cross Validation,K-CV)寻优SVR参数,得到预测模型。最后,预测了2018~2020年海南支线航空客运市场需求,从而为其建设支线机场提供一定的决策参考和可靠的理论依据,具有一定的现实意义和应用价值。For predicting the market demand of regional air transportation for passengers,some regions(such as Hainan)lack enough available data to establish accurate prediction models.This paper proposes a method based on Clustering and Support Vector Regression(SVR)to predict the market demand of air transportation for passengers in such regions.Firstly,the paper finds the similar regions to Hainan in distribution ratio of the market demand,then compares Hainan with these regions which were clustered together.Secondly,the paper selects the data samples of the similar regions and forms a prediction model after getting the SVR parameters by K-fold Cross Validation(K-CV).Finally,the paper predicts the market demand of Hainan air transportation for passengers from 2018 to 2020.The results could provide theoretical support and guidance for constructing new regional airports,which is realistic and practical value to some extent.

关 键 词：支线航空 客运市场需求 预测 系统聚类 SVR 

分 类 号：TPTP18[自动化与计算机技术]