Table 3.

A comparison of uncertainty measures of the DBN (section 6), static BN [15] and fuzzy logic assessment systems [14] expressed as a histogram of the ratio of automatic quality metrics and manual error bars across a test-set of 2,672 samples for the temperature sensors in Sullivans Cove. (a) Temperature sensor situated at a depth of 1 m; (b) Temperature sensor situated at a depth of 10 m.

Automatic EB / Manual EB	DBN		static BN		fuzzy logic
	instances	%	instances	%	instances	%
Above 300 %	13	0.5	16	0.6	13	0.5
200–300 %	57	2.1	55	2.1	166	6.2
150–200 %	36	1.4	117	4.4	119	4.5
66–150 %	2,488	93.1	2,374	88.9	2,131	79.8
50–66 %	42	1.6	56	2.1	108	4.0
33–50 %	28	1.0	36	1.4	92	3.4
Less than 33%	8	0.3	18	0.6	43	1.6
(a)

Automatic EB / Manual EB	DBN		static BN		fuzzy logic
	instances	%	instances	%	instances	%
Above 300 %	9	0.3	8	0.3	1	0.0
200–300 %	27	1.0	50	1.9	339	12.7
150–200 %	11	0.4	113	4.2	84	3.1
66–150 %	2,592	97.0	2421	90.6	2,224	83.2
50–66 %	24	0.9	52	1.95	9	0.3
33–50 %	6	0.2	14	0.5	7	0.3
Less than 33%	3	0.1	14	0.5	8	0.3
(b)