发表刊物：Applied Geography

关键字：Intercity travel; Machine learning; Mode choice; Nonlinear effect; XGBoost

摘要：Unraveling the complex relationships between intercity travel mode choices and their determinants is valuable for transport planning and development. However, a better understanding of the key influential factors shaping passengers' comprehensive mode choices for intercity travel, as well as the interactive and nonlinear effect between these two, is still needed. By conducting field surveys in Xi'an, China, a tourist city, this study collected intercity travel mode data for airplanes, high-speed railway (HSR), traditional trains, and express buses, as well as a comprehensive spectrum of passengers' attribute factors, including relatively underexplored factors such as online ticketing methods. The XGBoost algorithm and SHAP analysis were applied to conduct an in-depth and interpretable effect diagnosis. The results reveal obvious interactions among passengers' socioeconomic and travel behavior attributes. Among them, we find that online ticketing methods play an indispensable role in different population groups. Additionally, travel distance has the most significant nonlinear effects on intercity multimodal choices; this demonstrates a highly heterogeneous influential effect and a higher sensitivity than for intra-city travels. This study provides new evidence to better understand intercity travel behaviors and provides policymakers with targeted strategies for regional and national intercity transport planning and management.

合写作者：石兰馨,师洋,汤俊卿,赵鹏军,王玉婷,陈君

第一作者：李晓伟

论文类型：期刊论文

文献类型：J

是否译文：否

发表时间：2024-01-01

收录刊物：SSCI