Spatial cluster detection using dynamic programming

Table 2 Additional model averaging comparison

Frequency range	Number of rectangles	DP AUC	SR AUC	(DP-SR) 95% confidenceinterval limits		p-value
Low	1	0.51545	0.51647	- 0.01450	0.01247	0.8830
	2	0.52214	0.52353	- 0.01462	0.01185	0.8373
	3	0.55150	0.55355	- 0.01652	0.01244	0.7824
	4	0.55416	0.55637	- 0.02003	0.01561	0.8080
Mid	1	0.66873	0.67618	- 0.01967	0.00478	0.2326
	2	0.80826	0.82422	- 0.02771	- 0.00420	0.0078
	3	0.78402	0.79984	- 0.03444	0.00650	0.0520
	4	0.88198	0.88272	- 0.01471	0.01324	0.9177
High	1	0.88433	0.91615	- 0.04173	- 0.02191	< 0.0001
	2	0.96696	0.97218	- 0.01147	0.00105	0.1027
	3	0.97819	0.98139	- 0.00749	0.00109	0.1438
	4	0.99626	0.99567	- 0.00224	0.00340	0.6852

Comparison DP and SR ROC curves obtained from model averaging for outbreaks composed of various combinations of the rectangles in Figure 9

ISSN: 1472-6947