A general framework for estimation and inference of geographically weighted regression models: 1. Location-specific kernel bandwidths and a test for locational heterogeneity.

The writers argue that placing geographically weighted regression (GWR) within a statistical context as a spatial model of error variance heterogeneity, or what might be called locational heterogeneity, solves the problems associated with developing GWR along the lines of local regression and smooth...

Descripción completa

Detalles Bibliográficos
Publicado en:Environment & Planning A Vol. 34; no. 4; pp. 733 - 755
Autores principales: Páez, Antonio, Uchida, Takashi, Miyamoto, Kazuaki
Formato: Artículo
Publicado: Pion Limited April 2002
Materias:
Acceso en línea:Ver este registro en EBSCOhost
Descripción
Sumario:The writers argue that placing geographically weighted regression (GWR) within a statistical context as a spatial model of error variance heterogeneity, or what might be called locational heterogeneity, solves the problems associated with developing GWR along the lines of local regression and smoothing techniques. They present a maximum-likelihood-based framework for estimation and inference of a general GWR model that leads to a method for estimating location-specific kernel bandwidths.