A modification to geographically weighted regression

Table 7 Average variances and biases of \(\beta_{ 0}\) and \(\beta_{1}\) on a linear surface (single-type)

S/N of \(\beta_{ 0}\):	OLS		Basic GWR		Local linear		CGWR
S/N of \(\beta_{ 0}\):	1	5	1	5	1	5	1	5
(i) \(\beta_{ 0}\)
Conditional on S/N of \(\beta_{1}\) = 1
Variance	0.009	0.011	0.039	0.067	0.051	0.052	0.024	0.036
Bias	0.029	0.028	0.002	0.004	0.0003	0.0002	0.001	0.0002
Conditional on S/N of \(\beta_{1}\) = 5
Variance	0.010	0.009	0.090	0.107	0.057	0.043	0.048	0.057
Bias	0.723	0.719	0.026	0.038	0.0005	0.0002	0.012	0.009
(ii) \(\beta_{1}\)
Conditional on S/N of \(\beta_{1}\) = 1
Variance	0.031	0.036	0.113	0.190	0.146	0.145	0.071	0.117
Bias	0.038	0.695	0.013	0.051	0.001	0.0005	0.004	0.028
Conditional on S/N of \(\beta_{1}\) = 5
Variance	0.030	0.027	0.248	0.294	0.152	0.114	0.088	0.152
Bias	0.191	0.880	0.048	0.124	0.001	0.0002	0.004	0.010

ISSN: 1476-072X