Patient Specific Seizure Prediction System Using Hilbert Spectrum and Bayesian Networks Classifiers

Table 2

Evaluation of the proposed seizure prediction system.

Patient; number	Number of; Sz.	Interictal; hours	CFS					No feature selection
			Sens.; (%)	Det. Lat.; (min)	FPs/h	FP%	value	Sens.; (%)	Det. Lat.; (min)	FPs/h	FP%	value
1	4	23.91	100	35	0	0	0	100	27.5	0	0	0
2	3	23.89	100	35	0.08	5.178	0.0001	66.67	27.5	0	0	0
3	5	23.73	100	34	0	0	0	100	29	0.08	5.296	0
4	5	23.91	100	33	0.04	2.500	0	100	33	0.04	2.500	0
5	5	23.64	100	27	0.51	35.000	0.0031	80	23.75	0.63	40.625	0.0067
6	3	23.56	100	35	0.42	33.443	0.0272	100	33.33	0.30	21.967	0.0075
7	3	24.50	100	35	0	0	0	100	35	0.16	10.760	0.0009
8	2	24.06	100	20	0.21	17.105	0.023	100	22.5	0.21	14.803	0.0172
9	5	23.84	100	35	0	0	0	100	35	0.04	2.477	0
10	5	24.35	100	35	0.21	16.109	0.0001	100	35	0.33	27.964	0.001
11	4	23.02	75	35	0	0	0	75	35	0.22	18.301	0.015
12	4	24.63	100	35	0.04	2.462	0	100	35	0.12	9.846	0.0001
13	2	23.78	50	35	0	0	0	0	—	0	0	1
14	4	23.30	100	31.25	0.39	25.566	0.0027	75	30	0.56	41.100	0.1485
15	4	23.75	100	35	0.08	6.349	0	100	35	0.25	17.143	0.0005
16	5	23.92	100	29	0.42	24.768	0.0005	100	31	0.42	24.768	0.0005
17	5	23.98	100	35	0.29	18.154	0.0001	100	35	0.38	21.231	0.0002
18	5	24.79	100	35	0.20	13.773	0	100	35	0.40	22.455	0.0003
19	4	24.25	75	35	0.82	53.125	0.2931	75	35	0.78	53.125	0.2931
20	5	24.79	100	35	0.52	32.537	0.0021	80	35	0.65	48.358	0.1224
21	5	23.85	100	31	0.04	4.334	0	100	35	0.25	20.433	0.0002
	87	503.45	96.55	33.21	0.21	13.896	0.000	90.8	32.53	0.28	19.347	0.000