建成环境对路内停车泊位使用率的非线性影响模型

Nonlinear model for impact of built environment on curb parking spaces occupancy

为精细化把握微观空间尺度下的路内停车需求规律,针对工作日和非工作日两种典型场景,分别构建了分位数回归模型,揭示了土地利用、停车站点属性、交通因素、社会人口和经济对路内停车泊位使用率的非线性影响。研究结果表明:分位数回归模型捕捉研究变量间复杂关系的性能显著优于传统线性回归模型;解释变量的回归系数和显著性水平随着分位数的变化而变化,且工作日和非工作日存在明显异质性。例如,停车站点周边公交站的数量与低、高泊位使用率密切相关,而该变量在中等泊位使用率下无统计学意义;工作日餐饮类兴趣点对路内停车泊位使用率的影响并不显著。

To finely grasp the curb parking demand pattern at a micro-spatial scale,quantile regression models for two typical scenarios,namely workdays and weekends were proposed,to explore the nonlinear effects of land utilization,curb parking spaces,traffic factors,and socioeconomic and population on curb parking space occupancy.The results show that quantile regression is superior to linear regression in capturing the complex relationship between variables.And the coefficients and significance levels of explanatory variables change as the quantile changes.At the same time,there is an obvious heterogeneity between workdays and weekends.For example,the number of bus stations around the parking sites is closely related to the low and high occupancies,but this variable has no statistical significance under the medium occupancy.The catering services POIs do not significantly influence curb parking space occupancy on workdays.