Modeling International Student Migrant Tables
Modeling International Student Migrant Tables
This paper demonstrates the use of spatial interaction models for international student migrant tables
using a negative binomial regression in order to account for overdispersion. The Expectation-
Maximization (EM) algorithm is used in fitting these models to account for missing cells, which are a
common occurrence in international population mobility tables. Data for the five largest sending and
receiving nations of international student migrants between 1998 and 2005 are used. The results of
fitting a quasi-independent model, main effects models with multiple covariates and interaction models
are compared with respect to the Akaike Information Criterion in order to establish the most
parsimonious model. By using the EM algorithm to determine parameters in these models provides
imputations for cell values previously unknown.
Southampton Statistical Sciences Research Institute, University of Southampton
Abel, Guy J.
d35b5069-3c52-4d13-a678-1684ae1fce1e
31 July 2008
Abel, Guy J.
d35b5069-3c52-4d13-a678-1684ae1fce1e
Abel, Guy J.
(2008)
Modeling International Student Migrant Tables
(S3RI Methodology Working Papers, M08/05)
Southampton, UK.
Southampton Statistical Sciences Research Institute, University of Southampton
19pp.
Record type:
Monograph
(Working Paper)
Abstract
This paper demonstrates the use of spatial interaction models for international student migrant tables
using a negative binomial regression in order to account for overdispersion. The Expectation-
Maximization (EM) algorithm is used in fitting these models to account for missing cells, which are a
common occurrence in international population mobility tables. Data for the five largest sending and
receiving nations of international student migrants between 1998 and 2005 are used. The results of
fitting a quasi-independent model, main effects models with multiple covariates and interaction models
are compared with respect to the Akaike Information Criterion in order to establish the most
parsimonious model. By using the EM algorithm to determine parameters in these models provides
imputations for cell values previously unknown.
Text
55485-01.pdf
- Author's Original
More information
Published date: 31 July 2008
Identifiers
Local EPrints ID: 55485
URI: http://eprints.soton.ac.uk/id/eprint/55485
PURE UUID: 81bd13af-f339-4583-aa51-6ee5f89c3993
Catalogue record
Date deposited: 31 Jul 2008
Last modified: 20 Feb 2024 03:22
Export record
Contributors
Author:
Guy J. Abel
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics