Destination Choice of Athenians: An Application of Geographically Weighted Versions of Standard and Zero Inflated Poisson Spatial Interaction Models

Abstract

The main aim of this article is to combine recent developments in spatial interaction modeling to better model and explain spatial decisions. The empirical study refers to migration decisions made by internal migrants from Athens, Greece. To achieve this, geographically weighted versions of standard and zero inflated Poisson (ZIP) spatial interaction models are defined and fit. In the absence of empirical studies for the effect of potential determinants on internal migration decisions in Greece and the presence of an excessive number of zero migration flows among municipalities in Greece, this article provides empirical evidence for the power of the proposed Geographically Weighted ZIP regression method to better explain destination choices of Athenian internal migrants. We also discuss statistical inference issues in relation to the application of the proposed regression techniques.

Geographical random forests: a spatial extension of the random forest algorithm to address spatial heterogeneity in remote sensing and population modelling

“…The area that the sub-model operates in is called the neighbourhood (or kernel), and the maximum distance between a data point and its kernel is called the bandwidth (Brunsdon et al 1998). There are two usual types of kernels, 'adaptive' and 'fixed' (Kalogirou 2015). In the former, the neighbourhood is defined by n nearest neighbours and in the latter, by a circle whose radius is the bandwidth (Brunsdon et al 1998;Fotheringham et al 2003).…”

Using Disaster Outcomes to Validate Components of Social Vulnerability to Floods: Flood Deaths and Property Damage across the USA

“…Residuals were significantly spatially autocorrelated for death model residuals (Moran's I = 0.282, p < 0.001). However, methods for implementing spatial weights for zero-inflated regression with a negative binomial distribution were not available at the time this paper was written (for a zero-inflated geographically weighted regression with a Poisson distribution, see the lctools package from [135]). Spatial autocorrelation urges caution in model interpretation, as the model fit and coefficient estimates could be overestimated.…”

The spatial structure debate in spatial interaction modeling: 50 years on

“…For example, it does not provide strong empirical evidence for any particular form, and it is not clear how each form would arise, as has been done for the CD framework. As a result the CD model has been applied in many domains including urban modeling (Fotheringham 1985;Fotheringham and Knudsen 1986), the study of telecommunications flows (Guldmann 1999), access to heath facilities (de Mello-Sampayo 2016), crime location analysis (Bernasco 2010), determinants of trade (de Mello-Sampayo 2017a,b), and recreation and tourism (Matthews et al 2018), though it has enjoyed particular popularity within migration modeling (Ishikawa 1987;Ishikawa 1990;Fik et al 1992;Pellegrini and Fotheringham 1999;Fotheringham et al 2000;Pellegrini and Fotheringham 2002;Yano et al 2003;Fotheringham et al 2004;Kalogirou 2015), commuting-to-work research (Thorsen and Gitlesen 1998;Gitlesen and Thorsen 2000;Gitlesen et al 2010), and retail analysis (Fotheringham and Knudsen 1986;Guy 1987;Fotheringham 1988;Pellegrini et al 1997;Birkin et al 2010). The underlying theory of the CD model has also been explored via simulation studies (Lo 1991a;Fotheringham et al 2001;Ubøe et al 2008) and it has been connected with other key concepts, such as central place theory (Fik and Mulligan 1990) and trip chaining (Bernardin et al 2009).…”