Work

An extension of geographically weighted regression with flexible bandwidths

Year: 2014
Type: dissertation
Abstract: Various statistical methods have been developed for local spatial analysis. Among \nthem Geographically Weighted Regression (GWR) is a simple yet powerful method \nto explore spatially varying relationships between variables. This thesis examines how \nGWR can be extended to investigate spatially varying relationships at various \ngeographical scales within one model. \nGWR assumes that observations near to a regression location have more influence on \nthe estimation of local regression coefficients than do observations farther away. A \nsingle bandwidth is employed in basic GWR to control the rate of distance-decay in \nthis influence. The magnitude of the bandwidth affects the scale of variation in the \nestimated regression coefficients and thus usefully reflects the appropriate spatial \nscale at which the processes being modelled operate. A small bandwidth suggests the \nprocesses operate over a local spatial scale, whilst a large bandwidth indicates a more \nregional process. \nIn practice, a single bandwidth as in basic GWR may not be sufficient to reflect the \npotentially complex spatial variations in relationships between variables in a \nmultivariate spatial model. Therefore, in order to estimate coefficient surfaces that \nmay vary at different spatial scales for different variables, Flexible Bandwidth GWR \n(FBGWR) is proposed to allow different bandwidths to be individually specified for \neach independent variable in a regression framework. An algorithm based on back- \nfitting is developed to calibrate the FBGWR model. \nThe performance of FBGWR is investigated with simulated datasets where \ncoefficients are predefined at various levels of non-stationarity across space. A case \nstudy is then carried out on data relating to the Irish Famine to demonstrate the \napplication of FBGWR to real-world processes. The results suggest that FBGWR can \ndistinguish various scales of non-stationarity in spatial processes and provide an \nimproved model over basic GWR. FBGWR therefore represents a useful development \nin the modelling of spatially varying processes.
Author: Wenbai Yang
Language: English
Open Access status: green
OpenAlex ID: W1250261183
Cites: 146
Cited by: 106
Related to: 10