An Advanced Broyden–Fletcher–Goldfarb–Shanno Algorithm for Prediction and Output-Related Fault Monitoring in Case of Outliers

Table 1

FDRs of the 21 faults in the TE benchmark (%).

Fault	PLS		BFGS		KICA		ABFGS
	1	0.00	0.00	50.75	88.63	98.75	98.75	17.88	99.88
2	0.00	0.00	94.00	92.25	98.63	98.50	96.75	98.13
5	5.50	5.38	6.25	6.13	20.50	21.38	8.25	100.00
6	30.25	28.88	79.00	79.50	97.25	98.25	95.63	100.00
7	6.75	6.88	9.00	9.38	37.88	25.00	18.13	100.00
8	32.50	32.13	38.75	40.13	73.63	87.50	39.38	98.13
10	6.88	6.63	10.75	10.63	30.88	9.50	2.25	84.63
12	21.38	20.25	22.38	22.50	68.38	61.75	58.00	99.88
13	41.38	41.00	44.00	43.63	88.63	81.50	61.25	95.25
16	0.25	0.25	1.13	1.00	5.38	1.75	1.25	89.25
17	5.00	5.00	9.50	10.00	7.38	0.75	62.75	96.88
18	30.75	28.88	75.00	75.25	82.38	86.00	84.25	89.88
20	15.13	14.88	16.75	18.75	10.38	5.00	6.75	89.88
21	10.63	10.88	29.38	27.75	36.75	18.50	0.38	43.38
3	4.13	4.13	4.00	4.00	10.63	2.38	0.13	0.38
4	0.00	0.00	1.50	1.50	4.25	0.63	22.00	100.00
9	0.50	0.50	0.38	0.25	0.63	0.25	1.38	2.63
11	1.25	1.25	2.25	2.38	14.25	0.38	10.88	71.38
14	8.13	8.25	9.75	10.13	0.88	2.38	38.63	100.00
15	1.63	1.63	2.13	2.00	6.13	0.88	4.50	2.00
19	1.00	1.00	2.38	2.13	2.25	1.75	2.25	90.13
AVG	10.62	10.37	24.24	26.09	37.89	33.46	30.13	78.65
AVG-FAR	0.63	0.63	0.54	0.57	0.95	1.04	0.86	0.74

The bold data in the table represent the highest fault detection rate.