Do the risk factors for type 2 diabetes mellitus vary by location? A spatial analysis of health insurance claims in Northeastern Germany using kernel density estimation and geographically weighted regression

Table 2 Comparison of bandwidth types, kernel shapes and bandwidth optimization methods

Modell (bandwidth type, kernel shape, optimization method)	AICc	Adjusted R²	Moran’s I of residuals
Adaptive, Gaussian, AICc	−347	0.51	p < 0.001
Adaptive, Gaussian, AIC	−347	0.51	p < 0.001
Adaptive, Gaussian, BIC	−315	0.44	p < 0.001
Adaptive, Gaussian, CV	−347	0.51	p < 0.001
Fixed, Gaussian, AICc	−385	0.62	p < 0.05
Fixed, Gaussian, AIC	−265	0.66	p > 0.05
Fixed, Gaussian, BIC	−316	0.44	p < 0.001
Fixed, Gaussian, CV	−370	0.64	p > 0.05
Adaptive, bi-square, AICc	−394	0.63	p < 0.001
Adaptive, bi-square, AIC	−374	0.66	p > 0.05
Adaptive, bi-square, BIC	−320	0.45	p < 0.001
Fixed, bi-square, AICc	−385	0.62	p < 0.01
Fixed, bi-square, AIC	40	0.68	p > 0.05
Fixed, bi-square, BIC	−316	0.44	p < 0.001

ISSN: 1476-072X