Predicting the distribution of households and employment: a seemingly unrelated regression model with two spatial processes

Abstract

ABSTRACT:Household and employment counts (by type) are key inputs to models of travel demand. For a variety of reasons, spatial dependence is very likely present in and across these counts. In order to identify the nature of these unobserved relationships, this study performs a series of Lagrange multiplier tests to confirm the co-existence of spatial lag and error processes within individual equations (6 household types and 3 employment categories). It then provides the first application of a feasible generalized spatial 3SLS estimation procedure for a seemingly unrelated regression (SUR) model of these equations.In the resulting model of Austin, Texas data, local land use conditions offer substantial predictive power of households and jobs, and transportation access plays a role, as anticipated. The work demonstrates that SUR estimation of land use intensities from parcel-level data with two types of spatial dependence is feasible and meaningful. Coupled with an upstream model of land use type, this work offers the key inputs for travel demand analyses, with transportation system performance feedback.

Proximity to Natural Amenities: A Seemingly Unrelated Hedonic Regression Model with Spatial Durbin and Spatial Error Processes

“…In the first case, a spatial lag model for the housing equation is estimated via a maximum likelihood approach (ML) and the LM test is calculated with the null hypothesis being λ house = 0 as outlined in and Bera (1998). In a similar manner, the LM test for spatial lag autocorrelation in the presence of spatial error autocorrelation is derived by first estimating a spatial error model (Anselin et al 1996; Zhou and Kockelman ). In this case, the null hypothesis is ρ = 0.…”

“…There is also the possibility to include both a spatial lag of the response variable and a spatially autocorrelated error term in the same model if there is no a priori spatial specification in mind (Ham et al ; Kelejian and Prucha ; Zhou and Kockelman ). While this specification appears to be more flexible than that of equations () or (5), it is does not allow for the possibility of including spatially lagged independent variables as the estimated parameters would then be unidentified (Manski ).…”

“…The goodness of fit measure used to determine the combination of weight matrices that best fit our dataset was the McElroy R 2 measure (McElroy ). Following the notation in Zhou and Kockelman (), this diagnostic statistics was calculated in the following way: where E[ ɛɛ ′ | X 1 , X 2 , …, X M ] = Ω, is the ij th element in the matrix , and is the mean value of the response variable (e.g., P or HHINC) for equation i .…”

Exploring time variants for short-term passenger flow

“…During the past two decades, a number of researchers have developed various travel demand forecasting models to predict passenger flow and trends, such as conventional travel demand modeling, and multiple regression (e.g., Alfa, 1986;Wirasinghe and Kumarage, 1998;Kulshreshtha and Nag, 2000;Golias, 2002;Jovicic and Hansen, 2003;Bar-Gera and Boyce, 2003;Varagouli et al, 2005;Wardman, 2006;Tsekeris and Stathopoulos, 2006;Zhou and Kockelman, 2009). The conventional travel demand forecasting model is a sequential demand modeling method considering trip generation, trip distribution, mode choice, and assignment modules.…”

Measuring the structural determinants of urban travel demand

“…We have verified that they are statistically significant and checked whether they have a similar impact on urban travel in different groups of countries. We have used a robust econometric method (2SLS, SUR, 3SLS 2 , Chow's stability test, see Greene, 1993;Maddala, 2008) which has, to the best of our knowledge, only occasionally been used in the sphere of transport (apart from by Cervero and al., 2002;Zhou and al., 2008). We have then checked whether our findings agree with those in the literature.…”