摘要
在空铁联运出行中,为旅客推荐合理的中转城市是用户行程规划中的关键一步。研究基于旅客选择的中转城市数据,运用特征选择与数据处理技术,通过对比支持向量机(SVM)、K最近邻算法(KNN)、随机森林算法(RF)、多层感知机(MLP)和XGBoost模型的实验结果,最后采用结果最优的XGBoost模型对空铁联运中转城市进行推荐。结果表明,该模型在正确率上达到91.92%,具有一定的先进性与优越性;对于采用数据增强后训练的XGBoost模型,大大增加了模型的鲁棒性。在影响因素分析中,结果表明出发与到达的客观属性信息比旅客个体特征贡献度更高,也是首次证明出发与到达的客观属性信息对旅客选择中转城市有重要影响,为空铁联运中转城市推荐研究提供了新的思路。
In the air-rail intermodal travel, it is a key step in the users’ travel planning to recommend a reasonable transit city for passengers. The study was based on the data of transit cities selected by passengers, used feature selection and data processing technology, and finally adopted the XGBoost model with the best results to recommend the transit cities of air-rail intermodal transport, by comparing the experimental results of Support Vector Machine(SVM), K-nearest Neighbor Algorithm(KNN), Random Forest Algorithm(RF), Multi-layer Perceptron(MLP) and XGBoost model. The results show that the accuracy of the model reaches up to 91.92%, which reflects some progressiveness and superiority.For XGBoost model trained after data augmentation, the robustness of the model is greatly increased. In the analysis of influencing factors, the results show that the objective attribute information of departure and arrival has a higher contribution than the individual characteristics of the passengers. This is also the first time to prove that the objective attribute information of departure and arrival has an important impact on the selection of transit cities by passengers,which casts new light on the research of the recommendation of transit cities for air-rail intermodal transport.
作者
白广栋
翁湦元
张启蒙
朱建军
黄家玮
BAI Guangdong;WENG Shengyuan;ZHANG Qimeng;ZHU Jianjun;HUANG Jiawei(Institute of Computing Technology,China Academy of Railway Sciences Corporation Limited,Beijing 100081,China)
出处
《铁道运输与经济》
北大核心
2023年第3期24-31,共8页
Railway Transport and Economy
基金
中国国家铁路集团有限公司科技研究开发计划课题(N2022S003)。
关键词
空铁联运
中转城市推荐
机器学习
数据增强
XGBoost模型
Air-Rail Intermodal Transport
Recommendation of Transit Cities
Machine Learning
Data Augmentation
XGBoost Model
作者简介
通信作者:翁湦元(1991-),男,福建福州人,中国铁道科学研究院集团有限公司电子计算技术研究所助理研究员。
