Specification tests and model selection for aggregate spatial interaction: an empirical comparison.

L. Anselin

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

Five simple theories on aggregate spatial interaction behavior were compared with respect to their performance on a set of observations on Canadian interprovincial movements. Three different tests on nonnested alternatives were applied to compare all models in a pairwise fashion. The Cox-Pesaran-Deaton test and the J-test, particularly suitable for linear alternatives, and the P test, computationally easier in the nonlinear case, though asymptotically equivalent, show different values in finite sample cases. Also, a test on model discrimination based on the Akaike Information Criterion was illustrated.-from AuthorCity & Regional Planning Dept, Ohio State Univ, USA.

Original languageEnglish (US)
Pages (from-to)1-15
Number of pages15
JournalJournal of Regional Science
Volume24
Issue number1
StatePublished - 1984
Externally publishedYes

Fingerprint

interaction
interaction behavior
Akaike information criterion
regional planning
discrimination
test
comparison
performance
Values

ASJC Scopus subject areas

  • Environmental Science (miscellaneous)
  • Development

Cite this

Specification tests and model selection for aggregate spatial interaction : an empirical comparison. / Anselin, L.

In: Journal of Regional Science, Vol. 24, No. 1, 1984, p. 1-15.

Research output: Contribution to journalArticle

@article{0b5b470b181e47dbbef5bf6ae1cffb9e,
title = "Specification tests and model selection for aggregate spatial interaction: an empirical comparison.",
abstract = "Five simple theories on aggregate spatial interaction behavior were compared with respect to their performance on a set of observations on Canadian interprovincial movements. Three different tests on nonnested alternatives were applied to compare all models in a pairwise fashion. The Cox-Pesaran-Deaton test and the J-test, particularly suitable for linear alternatives, and the P test, computationally easier in the nonlinear case, though asymptotically equivalent, show different values in finite sample cases. Also, a test on model discrimination based on the Akaike Information Criterion was illustrated.-from AuthorCity & Regional Planning Dept, Ohio State Univ, USA.",
author = "L. Anselin",
year = "1984",
language = "English (US)",
volume = "24",
pages = "1--15",
journal = "Journal of Regional Science",
issn = "0022-4146",
publisher = "Wiley-Blackwell",
number = "1",

}

TY - JOUR

T1 - Specification tests and model selection for aggregate spatial interaction

T2 - an empirical comparison.

AU - Anselin, L.

PY - 1984

Y1 - 1984

N2 - Five simple theories on aggregate spatial interaction behavior were compared with respect to their performance on a set of observations on Canadian interprovincial movements. Three different tests on nonnested alternatives were applied to compare all models in a pairwise fashion. The Cox-Pesaran-Deaton test and the J-test, particularly suitable for linear alternatives, and the P test, computationally easier in the nonlinear case, though asymptotically equivalent, show different values in finite sample cases. Also, a test on model discrimination based on the Akaike Information Criterion was illustrated.-from AuthorCity & Regional Planning Dept, Ohio State Univ, USA.

AB - Five simple theories on aggregate spatial interaction behavior were compared with respect to their performance on a set of observations on Canadian interprovincial movements. Three different tests on nonnested alternatives were applied to compare all models in a pairwise fashion. The Cox-Pesaran-Deaton test and the J-test, particularly suitable for linear alternatives, and the P test, computationally easier in the nonlinear case, though asymptotically equivalent, show different values in finite sample cases. Also, a test on model discrimination based on the Akaike Information Criterion was illustrated.-from AuthorCity & Regional Planning Dept, Ohio State Univ, USA.

UR - http://www.scopus.com/inward/record.url?scp=0021331495&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0021331495&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:0021331495

VL - 24

SP - 1

EP - 15

JO - Journal of Regional Science

JF - Journal of Regional Science

SN - 0022-4146

IS - 1

ER -