摘要
传统线性模型无法显示指标的非线性结果,难以体现各指标的相对重要性和影响阈值。利用深圳轨道交通刷卡数据,构建梯度提升决策树(GBDT)模型,探索建成环境对不同时间段客流量的非线性关系,以期为城市尤其是超大特大城市的交通规划、站点开发、站点周边建成环境优化提供参考。结果表明:不同建成环境对不同时间段客流量的相对重要性不一致,建筑容积率、居住用地比例、商业商务用地比例、距CBD距离对站点客流量均有相对较高且稳健的影响;建成环境与客流量存在明显的非线性关系和阈值;当土地利用混合度在0.60以上和距CBD距离超过17 km时,全天客流量呈现明显的下降趋势;建成环境对不同时间段客流量的影响机制不完全相同;城中村对地铁客流量的促进效果十分显著。
Traditional linear models cannot display the nonlinear results of indicators,making it difficult to reflect the relative importance and impact threshold of each indicator.This study uses Shenzhen rail transit smart card date to construct a gradient boosting decision tree(GBDT)model to explore the non-linear relationship between the built environment and passenger flow in different time periods,hoping to provide references for transportation planning,station development,and optimization of built environment around stations in cities,especially mega cities.The results show that:①Different built environments have different relative importance to passenger flow in different time periods.Building floor area ratio,proportion of residential land,proportion of commercial land,and distance from CBD all have relatively high and robust effects on site passenger flow.②There is a significant nonlinear relationship and threshold between the built environment and passenger flow.③When the land use mixing degree is above 0.60 and the distance from CBD exceeds 17 km,the daily passenger flow shows a significant downward trend.④The impact mechanism of the built environment on the passenger flow in different time periods is not completely the same.⑤The promotion effect of urban villages on subway passenger flow is very significant.
作者
黄建欣
龚蔚霞
张金林
汤咏诗
吴巍
HUANG Jianxin;GONG Weixia;ZHANG Jinlin;TANG Yongshi;WU Wei
出处
《上海城市规划》
北大核心
2023年第6期47-54,共8页
Shanghai Urban Planning Review
关键词
城市轨道交通
建成环境
客流量
梯度提升决策树
非线性影响
urban rail transit
built environment
passenger flow
gradient boosting decision tree
non-linear influencing
