Spatial Dependence in Trip Generation Models: A Spatial Econometric Perspective

Abstract

In most of the urbanized cities, socio-economic attributes tend to cluster as patterns of similarity in space, namely spatial effect, by agglomeration forces. Because of this effect, the classical linear regression model cannot sufficiently represent trip generation phenomenon. In order to correct this problem, we need a model which adequately deals with the spatial dependence patterns; that is the spatial economic model. This study identified the existence of spatial effect in the production and attraction of non-home base trips and developed the spatial econometric models for various trip purposes, using travel data of Daegu Metropolitan Area in South Korea. The results in this study indicated that NHBO production and attraction have the spatial error dependence and NHBO attraction also has the spatial lag dependence. In order to verify the results from models, supplementary efficacy tests to spatial clustering of NHBO attraction are performed, which shows the existence of both spatial dependence patterns. Consequently, it is concluded that the spatial econometric model is suitable for trip generation analysis.