近代回归分析在交通调查分析建模中的应用

本文以出行发生量模型的建立为例,较为系统地讨论了近代回归分析中的自变量选择及回归诊断方法在交通调查分析建模中的应用。分析了在获取调查数据后,如何借助回归自变量选择方法来选择最佳自变量子集,以确定简捷的回归模型;文中应用回归诊断方法,讨论了修正回归模型、探测错误的调查数据的方法,从而为建立简捷、高精度的交通模型打下了基础。最后,作者提出了应用自变量选择及回归诊断方法建立出行发行量模型的一般步骤。

Studies on the Prediction Model of People Trip Flow in Urban Aera

This thesis is to study the methods of establishing Trip Production Model by usingregression analysis. Some problems on the selection of independent variables and regressiondiagnostics have been discussed. Two examples are used to compare alls election rules,and the results show consistance. The influences on the results deriving from the overinfluence points, high leverage points and outliers are studied. The general procedure forthe selection of independent variables is shown in thesis, also the applicable methods ofpolynomial regression diagnostics, especially, the methods for detecting outliers. Whendependent variable is affected by some element which couldn't be quantified, we can addthe quanlitative variable to improve the results of regression. According to these studies,a method establishing Trip Production Model by using regression analysis has been suggest-ed in this thesis, which was used to set up the trip production model of Beijing, andthe result is satisfactory.