Understanding ridesplitting behavior of on-demand ride services: An ensemble learning approach

详细信息    查看全文

• 作者：Xiqun (Michael) Chen ; ^{chenxiqun@zju.edu.cn} ; Majid Zahiri ; Shuaichao Zhang
• 关键词：On-demand ride service ; Rideplitting ; Ensemble learning ; Boosting ; Decision tree
• 刊名：Transportation Research Part C: Emerging Technologies
• 出版年：2017
• 出版时间：March 2017
• 年：2017
• 卷：76
• 期：Complete
• 页码：51-70
• 全文大小：3509 K
• 卷排序：76

文摘

Present ensemble learning for better understanding ridesplitting behavior. Boosting trees improve prediction, generalizability, and robustness over base classifier. Explore on-demand ride services of DiDi (Taxi, Express, Private Car Service, and Hitch). Important features are ranked and selected by the ReliefF algorithm. Ensemble learning outperforms three popular classifiers in ridesplitting.