基于MNL模型的历史街区出行方式研究  

Travel Mode in Historical Street Based on MNL Model

作　　者：刘梦瑶 贺玉龙 孙小端 

机构地区：[1]北京工业大学城市交通学院,北京100124

出　　处：《交通运输研究》2017年第5期8-13,共6页Transport Research

基　　金："十二五"国家科技支撑计划项目(2014BAG01B01)

摘　　要：为了改善历史文化街区的交通出行环境,在北京南锣鼓巷胡同区及附近人口密集的区域开展调查,进行历史街区出行行为研究。首先,为了提高模型的准确度与可靠性,研究了不同服务人群、服务范围之间的差异性。通过对数据的初步分析,将出行方式划分为公共交通、慢行交通、私人交通3类,并分析得出该区域内慢行交通出行比例最低。然后,利用SPSS软件对特性变量与出行方式选择进行相关性分析,筛选特征变量。最后,建立多项Logit模型(Multinomial Logit Model,简称MNL)对历史街区出行行为进行分析,并使用南锣鼓巷实测数据对模型进行标定。研究结果表明,居民户口、年龄、月收入、汽车拥有量、胡同宽度、胡同单行线、游客影响、慢行交通意愿等因素对出行方式选择均有显著影响。In order to improve the traffic environment in the historical and cultural neighborhoods, aninvestigation in Nanluoguxiang and the densely populated areas nearby was conducted. The travel behav-iors of residents of Nanluoguxiang were studied. Firstly, in order to improve the accuracy and reliabilityof the model, the differences in service crowds and service areas were studied. Through the preliminaryanalysis of the data, the travel modes were divided into public transport, slow-moving traffic and privatetransport. The ratio of slow-moving traffic is the lowest among the three travel modes. Then, the SPSSwas used to analyze the relativity between attribute variables and travel mode selection, and the featurevariables were screened. Finally, the MNL(Multinomial Logit Model) for analyzing the travel behaviors ofresidents in the historical district was established, and the parameters were calibrated by the actual sur-vey data of Nanluoguxiang. The results show that residents′ household registration, age, monthly income,car ownership, alley′s width, alley′s single line, tourists′ influence and slow-moving traffic intentionshave significant influence on travel behavior.

关 键 词：历史街区 出行行为 出行环境 影响因素 多项Logit模型 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]