摘要
为高效准确预测旅客选择空铁联运各中转城市的概率并揭示相关因素影响机制,基于互联网客票脱敏数据,应用随机森林算法、特征重要度和部分依赖图方法探究解释变量与中转城市备选集间的复杂作用关系.结果表明,随机森林模型可有效处理不均衡分布样本,具有更高的预测精度,总体分类准确率可达88.54%,并具备描述自变量非线性作用的能力.以京津冀城市群保定上海的空铁联运场景为例,联运服务在运营方层面关于时间、价格、衔接效率等属性相较于旅客个人社会属性占据更高影响权重,且与中转城市选择概率间存在非线性联系,表现为Z形和S形曲线特征.该研究结果有助于空铁联运中转城市的优势市场划分,提升旅客联运服务质量.
To efficiently and accurately predict the probability of passengers choosing each transit city for air-rail intermodal travel and reveal the influence mechanism of related factors,based on Internet ticket-purchasing desensitization data,random forest algorithm,feature importance,and partial dependence plot methods were used to explore the complex interactions between explanatory variables and candidate choice sets of transit cities.The results show that the random forest model can effectively handle unevenly distributed samples,with higher prediction accuracy and overall classification accuracy up to 88.54%,thus it has ability to describe the nonlinear impacts of independent variables.Taking the scenario,air-rail intermodal travel from Baoding to Shanghai in Beijing-Tianjin-Hebei urban agglomeration as an example,the factors about intermodal travel service at operating level including time,price,and transferring efficiency have a higher weight than passengers'socioeconomic factors.There is a nonlinear relationship with the choice probabilities of transit cities,which are characterized by Z-shaped and S-shaped curves.The results are helpful for the advantageous market division of transit cities and improving service qualities for passengers in air-rail intermodal travel area.
作者
杨敏
任怡凤
盛强
刘冬梅
李宏伟
Yang Min;Ren Yifeng;Sheng Qiang;Liu Dongmei;Li Hongwei(School of Transportation,Southeast University,Nanjing 210096,China;Shanghai Ctrip Commerce Co.,Ltd.,Shanghai 200335,China;Research Institute of Highway Ministry of Transport,Beijing 100088,China)
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2022年第1期162-171,共10页
Journal of Southeast University：Natural Science Edition
基金
国家重点研发计划资助项目(2018YFB1601300)
国家自然科学基金资助项目(52072066)
江苏省自然科学基金杰出青年基金资助项目(BK20200014)
江苏省交通运输科技资助项目(2020Y12).
关键词
空铁联运
中转城市选择
随机森林模型
机器学习
air-rail intermodal travel
transit city choice
random forest algorithm
machine learning
