Spatial cluster detection using dynamic programming

Table 5 Comparison of model selection in terms of spatial recall

Frequency range	Coverage	DP recall	GR recall	SR recall	(DP-GR) 95% confidenceinterval limits		p-value
Low	< 16%	0.08729	0.09759	0.04993	- 0.01031	- 0.01029	< 0.00001
	16-34%	0.13665	0.14801	0.10795	- 0.01137	- 0.01135	< 0.00001
	≥ 35%	0.18579	0.19636	0.17592	- 0.01060	- 0.01055	< 0.00001
Mid	< 16%	0.39280	0.38533	0.11793	0.00745	0.00749	< 0.00001
	16-34%	0.56118	0.55025	0.24091	0.01092	0.01095	< 0.00001
	≥ 35%	0.74869	0.75813	0.42978	- 0.00956	- 0.00932	< 0.00001
High	< 16%	0.62454	0.60666	0.13601	0.01786	0.01789	< 0.00001
	16-34%	0.82999	0.82147	0.29433	0.00850	0.00852	< 0.00001
	≥ 35%	0.91850	0.91130	0.50550	0.00716	0.00724	< 0.00001
Very high	< 16%	0.93248	0.92970	0.24123	0.00277	0.00278	< 0.00001
	16-34%	0.99829	0.99809	0.62275	0.00020	0.00020	< 0.00001
	≥ 35%	0.99813	0.99815	0.74743	- 0.00002	- 0.00001	< 0.00001

Comparison of the DP, GR, and SR spatial recall obtained from model selection. The table shows the mean recall for each algorithm, the lower and upper 95% confidence limits for the difference in recall (DP-GR) and the associated t-test p-values

ISSN: 1472-6947