GEM-CRAP: a fusion architecture for focal seizure detection

Table 5 Model performance without pre-policy network activation

Class	Lable	Accuracy	Recall	F1-Score	Precision (%)
Subset4	0	0.94 ± 0.004	0.90 ± 0.001	0.92 ± 0.002	87.9 ± 0.4
	1	0.88 ± 0.002	0.74 ± 0.002	0.80 ± 0.003
	2	0.61 ± 0.005	0.91 ± 0.001	0.73 ± 0.004
Subset3	0	0.95 ± 0.005	0.91 ± 0.001	0.93 ± 0.003	89.5 ± 0.2
	1	0.86 ± 0.001	0.82 ± 0.001	0.84 ± 0.001
	2	0.59 ± 0.001	0.90 ± 0.004	0.71 ± 0.005
Subset5	0	0.88 ± 0.003	0.95 ± 0.005	0.91 ± 0.004	85.9 ± 0.1
	1	0.84 ± 0.001	0.60 ± 0.004	0.70 ± 0.002
	2	0.59 ± 0.001	0.38 ± 0.007	0.47 ± 0.003

ISSN: 1479-5876