基于有序Logistic回归的城市人行道服务水平研究

Pedestrian LOS of Urban Sidewalks Based on Orderly Logistic Regression

科学的人行道服务水平评价方法可以为营造良好的步行出行环境提供有力的理论支撑.针对人行道服务水平等级为定序变量的特点,本文提出了基于有序Logistic回归模型的人行道服务水平评价方法.首先,选择北京市交通枢纽地区34条典型人行道进行行人满意度调查,获取典型、全面的数据样本;然后,采用模糊C均值聚类方法得到行人满意度与人行道服务水平等级的对应关系;最后,应用逐步回归方法提取人行道服务水平显著性影响因素并分析其影响机理,进而建立人行道服务水平等级的有序Logistic回归模型.通过实际验证,模型精度较现有的线性回归模型有所提高.

Scientific method measuring pedestrian level of service（LOS） on sidewalks can provide a powerful theoretical support for a comfort walking trip environment. Aiming at the characteristics that pedestrian LOS is a orderly discrete variable, this paper proposed an evaluation method for pedestrian LOS, based on the orderly Logistic regression model. First, pedestrians＇ satisfaction questionnaire survey was conducted on34 representative sidewalks in Beijing transportation hub areas. Typical and comprehensive data were obtained. Then fuzzy C mean clustering was used to get the correspondence between pedestrians＇ satisfaction and the LOS. Finally, the significant influencing factors were extracted with step-regression method and their influencing mechanism was analyzed. And then the orderly Logistic regression model for pedestrian LOS was formulated. The field tast proved that the model has higher accuracy than the traditional linear model.