|
Abstract:
Geographically Weighted Regression has now become a popular technique in spatial analysis to uncover patterns of spatial nonstationarity in spatial processes. This talk will cover six recent refinements to the basic technique which improve or extend its use. These are: (i) semi-parametric GWR models with an empirical demonstration of its inclusion in new software GWR 4.0; (ii) model selection procedures for GWR; (iii) a solution to the multiple hypotheses testing problem; (iv) multicollinearity issues; (v) the use of different distance metrics; and (vi) the incorporation of fully flexible bandwidths.
|
