摘要
实际调查RP数据(Revealed preference data)具有可靠性,而意向调查SP数据(Stated Preference data)具有灵活性。为了更好地将两类数据结合起来,减少交通调查的工作量,提高调查的精确性,本文分析研究了RP/SP组和调查方案的设计,应用非集计模理论及调查所得石河子市组合数据,建立了交通方式划分的MNL模型,并应用spss13.0软件对模型参数进行了标定。通过模型拟合度检验验证了模型的有效性,和利用该模型对石河子市交通方式分担率进行了预测。结果表明,该模型能较全面地考虑居民出行选择的各方面影响因素,模型的预测精度和实用性较好。
RP data were more reliable, while SP data were more flexible. In order to join the two types of data together to reduce the traffic survey work volume and improve the accuracy of the survey, and based on the analysis of the design of RP/SP survey scheme, the application disaggregate model theory and the survey data of Shihezi city, amultinomial logit （MNL） model for predicting the traffic mode choice was developed, and spss13.0 was used to calibrate its parameters. After testifying the validity of the model by fitting of the model, the article takes the model to forecast the traffic share situation of Shihezi city. Experimental results show that the MNL model combined the RP/SP data can consider more factors affecting the resident travel mode choice, and can help to improve the prediction accuracy and practicability of the model.
出处
《交通运输工程与信息学报》
2008年第3期59-64,共6页
Journal of Transportation Engineering and Information
基金
"十一五"国家科技支撑计划项目(编号:2006BAJ18B03)
