Accommodating spatial correlation across choice alternatives in discrete choice models: An application to modeling residential location choice behavior

Ipek N. Sener, Ram Pendyala, Chandra R. Bhat

Research output: Contribution to journalArticle

37 Citations (Scopus)

Abstract

This paper presents a modeling methodology capable of accounting for spatial correlation across choice alternatives in discrete choice modeling applications. Many location choice (e.g., residential location, workplace location, destination location) modeling contexts involve choice sets where alternatives are spatially correlated with one another due to unobserved factors. In the presence of such spatial correlation, traditional discrete choice modeling methods that are often based on the assumption of independence among choice alternatives are not appropriate. In this paper, a Generalized Spatially Correlated Logit (GSCL) model that allows one to represent the degree of spatial correlation as a function of a multi-dimensional vector of attributes characterizing each pair of location choice alternatives is formulated and presented. The formulation of the GSCL model allows one to accommodate alternative correlation mechanisms rather than pre-imposing restrictive correlation assumptions on the location choice alternatives. The model is applied to the analysis of residential location choice behavior using a sample of households drawn from the 2000 San Francisco Bay Area Travel Survey (BATS) data set. Model estimation results obtained from the GSCL are compared against those obtained using the standard multinomial logit (MNL) model and the spatially correlated logit (SCL) model where only correlations across neighboring (or adjacent) alternatives are accommodated. Model findings suggest that there is significant spatial correlation across alternatives that do not share a common boundary, and that the GSCL offers the ability to more accurately capture spatial location choice behavior.

Original languageEnglish (US)
Pages (from-to)624-633
Number of pages10
JournalJournal of Transport Geography
Volume18
Issue number5
DOIs
StatePublished - Sep 2010

Fingerprint

residential location
modeling
choice of location
workplace
travel
methodology
ability

Keywords

  • Activity-travel behavior modeling
  • Discrete choice modeling
  • Distance-decay function
  • Residential location choice
  • Spatial correlation
  • Spatially correlated logit model

ASJC Scopus subject areas

  • Environmental Science(all)
  • Geography, Planning and Development
  • Transportation

Cite this

@article{55674a98ba674c8d8bf0a6b57377ec87,
title = "Accommodating spatial correlation across choice alternatives in discrete choice models: An application to modeling residential location choice behavior",
abstract = "This paper presents a modeling methodology capable of accounting for spatial correlation across choice alternatives in discrete choice modeling applications. Many location choice (e.g., residential location, workplace location, destination location) modeling contexts involve choice sets where alternatives are spatially correlated with one another due to unobserved factors. In the presence of such spatial correlation, traditional discrete choice modeling methods that are often based on the assumption of independence among choice alternatives are not appropriate. In this paper, a Generalized Spatially Correlated Logit (GSCL) model that allows one to represent the degree of spatial correlation as a function of a multi-dimensional vector of attributes characterizing each pair of location choice alternatives is formulated and presented. The formulation of the GSCL model allows one to accommodate alternative correlation mechanisms rather than pre-imposing restrictive correlation assumptions on the location choice alternatives. The model is applied to the analysis of residential location choice behavior using a sample of households drawn from the 2000 San Francisco Bay Area Travel Survey (BATS) data set. Model estimation results obtained from the GSCL are compared against those obtained using the standard multinomial logit (MNL) model and the spatially correlated logit (SCL) model where only correlations across neighboring (or adjacent) alternatives are accommodated. Model findings suggest that there is significant spatial correlation across alternatives that do not share a common boundary, and that the GSCL offers the ability to more accurately capture spatial location choice behavior.",
keywords = "Activity-travel behavior modeling, Discrete choice modeling, Distance-decay function, Residential location choice, Spatial correlation, Spatially correlated logit model",
author = "Sener, {Ipek N.} and Ram Pendyala and Bhat, {Chandra R.}",
year = "2010",
month = "9",
doi = "10.1016/j.jtrangeo.2010.03.013",
language = "English (US)",
volume = "18",
pages = "624--633",
journal = "Journal of Transport Geography",
issn = "0966-6923",
publisher = "Elsevier BV",
number = "5",

}

TY - JOUR

T1 - Accommodating spatial correlation across choice alternatives in discrete choice models

T2 - An application to modeling residential location choice behavior

AU - Sener, Ipek N.

AU - Pendyala, Ram

AU - Bhat, Chandra R.

PY - 2010/9

Y1 - 2010/9

N2 - This paper presents a modeling methodology capable of accounting for spatial correlation across choice alternatives in discrete choice modeling applications. Many location choice (e.g., residential location, workplace location, destination location) modeling contexts involve choice sets where alternatives are spatially correlated with one another due to unobserved factors. In the presence of such spatial correlation, traditional discrete choice modeling methods that are often based on the assumption of independence among choice alternatives are not appropriate. In this paper, a Generalized Spatially Correlated Logit (GSCL) model that allows one to represent the degree of spatial correlation as a function of a multi-dimensional vector of attributes characterizing each pair of location choice alternatives is formulated and presented. The formulation of the GSCL model allows one to accommodate alternative correlation mechanisms rather than pre-imposing restrictive correlation assumptions on the location choice alternatives. The model is applied to the analysis of residential location choice behavior using a sample of households drawn from the 2000 San Francisco Bay Area Travel Survey (BATS) data set. Model estimation results obtained from the GSCL are compared against those obtained using the standard multinomial logit (MNL) model and the spatially correlated logit (SCL) model where only correlations across neighboring (or adjacent) alternatives are accommodated. Model findings suggest that there is significant spatial correlation across alternatives that do not share a common boundary, and that the GSCL offers the ability to more accurately capture spatial location choice behavior.

AB - This paper presents a modeling methodology capable of accounting for spatial correlation across choice alternatives in discrete choice modeling applications. Many location choice (e.g., residential location, workplace location, destination location) modeling contexts involve choice sets where alternatives are spatially correlated with one another due to unobserved factors. In the presence of such spatial correlation, traditional discrete choice modeling methods that are often based on the assumption of independence among choice alternatives are not appropriate. In this paper, a Generalized Spatially Correlated Logit (GSCL) model that allows one to represent the degree of spatial correlation as a function of a multi-dimensional vector of attributes characterizing each pair of location choice alternatives is formulated and presented. The formulation of the GSCL model allows one to accommodate alternative correlation mechanisms rather than pre-imposing restrictive correlation assumptions on the location choice alternatives. The model is applied to the analysis of residential location choice behavior using a sample of households drawn from the 2000 San Francisco Bay Area Travel Survey (BATS) data set. Model estimation results obtained from the GSCL are compared against those obtained using the standard multinomial logit (MNL) model and the spatially correlated logit (SCL) model where only correlations across neighboring (or adjacent) alternatives are accommodated. Model findings suggest that there is significant spatial correlation across alternatives that do not share a common boundary, and that the GSCL offers the ability to more accurately capture spatial location choice behavior.

KW - Activity-travel behavior modeling

KW - Discrete choice modeling

KW - Distance-decay function

KW - Residential location choice

KW - Spatial correlation

KW - Spatially correlated logit model

UR - http://www.scopus.com/inward/record.url?scp=84857308488&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84857308488&partnerID=8YFLogxK

U2 - 10.1016/j.jtrangeo.2010.03.013

DO - 10.1016/j.jtrangeo.2010.03.013

M3 - Article

AN - SCOPUS:79953812453

VL - 18

SP - 624

EP - 633

JO - Journal of Transport Geography

JF - Journal of Transport Geography

SN - 0966-6923

IS - 5

ER -