Travel demand forecasts improved by using cross-sectional data from multiple time points: enhancing their quality by linkage to gross domestic product
Travel demand forecasts improved by using cross-sectional data from multiple time points: enhancing their quality by linkage to gross domestic product
复制标题
通过使用多个时间点的横截面数据改进旅行需求预测：通过与国内生产总值联系来提高其质量
DOI：
10.1007/s11116-016-9755-x
复制
发表时间：
2018
期刊：
影响因子：
4.3
通讯作者：
三古展弘
中科院分区：
文献类型：
--
作者：
Ryoji Ito and Kenneth C. Gehrt;Ryoji Ito;Ryoji Ito and Kenneth C. Gehrt;伊藤龍史;Ryoji Ito and Kenneth C. Gehrt;Ryoji Ito;Ichiro Furukawa;古川一郎・奥谷孝司;古川一郎;金春姫・古川一郎;古川 一郎;三古展弘
使用分类旅行需求模型的预测通常基于最近时间点的数据，即使可以从多个时间点获得截面数据。然而，这并不是对数据的很好利用。在之前的研究中，作者提出了一种联合利用多个时间点的截面数据的方法，其中参数假设为时间（年）的函数，即参数值随时间而变化。将该方法应用于日本名古屋的通勤模式选择分析。对2001年的行为进行了预测，其中一个模型仅使用了最近的1991年数据集，其他模型则结合了1971年、1981年和1991年的数据集。后一种模型的性能优于前一种模型，证明了所提方法的适用性。虽然时间函数将参数变化归因于随时间变化的趋势，但其理论基础需要进一步研究。本研究的目的是分析作者之前研究中使用的相同数据集，但将参数表示为人均国内生产总值（GDP）的函数。这种方法的理论基础问题较少，因为参数的变化是由经济条件的影响来解释的。人均GDP函数比时间函数给出了更好的预测。此外，在选择函数形式和外推到遥远的未来/过去甚至其他领域时，人均国内生产总值的函数出现的问题较少。对未来人均GDP不确定性的敏感性分析表明，所提出的模型是可行的。
Forecasts using disaggregate travel demand models are often based on data from the most recent time point, even when cross-sectional data is available from multiple time points. However, this is not a good use of the data. In a previous study, the author proposed a method that jointly utilises cross-sectional data from multiple time points in which parameters are assumed to be functions of time (year), meaning that the parameter values vary over time. The method was applied to journey-to-work mode choice analyses for Nagoya, Japan. Behaviours in 2001 were forecast using one model with only the most recent 1991 dataset and other models that combined datasets for 1971, 1981, and 1991. The latter models outperformed the former model, which demonstrated the applicability of the proposed method. Although the functions of time ascribe the parameter changes to the trends over time, the theoretical underpinnings needed further investigation. The aim of this study is to analyse the same dataset used in the author’s previous study, but express the parameters as functions of gross domestic product (GDP) per capita. This method has fewer problems related to its theoretical underpinnings, since the parameter changes are explained by the effects of economic conditions. The functions of GDP per capita produced better forecasts than the functions of time. In addition, the functions of GDP per capita present fewer problems when choosing functional forms and extrapolating into the distant future/past and even to other areas. Sensitivity analysis with respect to uncertainties in the future GDP per capita showed that the proposed models are practical.