Trust in Intrusion Detection Systems: An Investigation of Performance Analysis for Machine Learning and Deep Learning Models

Table 13

The performance of deep learning for WSN dataset using LSTM and GRU with selected features.

Model	Evaluation metric	Cross-validation performance					Testing performance
		Normal	Grayhole	Blackhole	Scheduling	Flooding	Normal	Grayhole	Blackhole	Scheduling	Flooding
LSTM with one layer	TN	98.59	99.91	99.48	96	99.93	98.50	99.92	99.73	96.61	99.99
	FP	1.41	0.09	0.52	4	0.07	1.50	0.08	0.27	3.39	0.01
	FN	6.82	8.5	39.19	0.47	9.9	1.63	5.19	40.12	0.33	9.22
	Accuracy	98.45	99.84	97.98	99.21	99.76	98.50	99.88	98.18	99.39	99.83
	Precision	64.84	90.6	83.37	99.59	96.08	64.42	91.49	89.92	99.66	99.47
	Recall	93.18	91.5	60.81	99.53	90.1	98.37	94.81	59.88	99.67	90.78
	F-score	76.2	91.02	69.98	99.56	92.98	77.85	93.12	71.89	99.66	94.93
LSTM with two layer	TN	98.41	99.9	99.17	94.62	99.92	98.48	99.91	99.25	94.51	99.88
	FP	1.59	0.1	0.83	5.38	0.08	1.52	0.09	0.75	5.49	0.12
	FN	3.3	13.3	42.91	0.83	21.43	5.10	4.11	42.50	0.80	19.88
	Accuracy	98.37	99.79	97.53	98.75	99.54	98.38	99.87	97.62	98.76	99.53
	Precision	62.68	89.59	73.95	99.45	94.85	63.25	90.43	75.66	99.44	92.62
	Recall	96.7	86.7	57.09	99.17	78.57	94.90	95.89	57.50	99.20	80.12
	F-score	76.04	87.68	64.37	99.31	85.85	75.91	93.08	65.34	99.32	85.92
GRU with one layer	TN	99.08	99.91	99.3	97.85	99.99	98.62	99.93	99.78	97.91	99.99
	FP	0.92	0.09	0.7	2.15	0.01	1.38	0.07	0.22	2.09	0.01
	FN	21.71	3.26	23.88	0.21	7.73	3.42	6.28	34.23	0.22	7.89
	Accuracy	98.52	99.88	98.4	99.61	99.85	98.57	99.87	98.46	99.60	99.85
	Precision	72.69	90.36	83.72	99.78	99.27	65.87	91.94	92.49	99.79	99.61
	Recall	78.29	96.74	76.12	99.79	92.27	96.58	93.72	65.77	99.78	92.11
	F-score	72.62	93.43	78.58	99.78	95.64	78.32	92.82	76.87	99.78	95.7
GRU with two layer	TN	98.53	99.93	99.25	89.1	99.81	98.65	99.91	99.49	94.11	99.82
	FP	1.47	0.07	0.75	10.9	0.19	1.35	0.09	0.51	5.89	0.18
	FN	5.85	39.26	53.04	0.54	19.99	7.44	9.78	42.09	0.49	13.73
	Accuracy	98.41	99.59	97.21	98.5	99.46	98.49	99.83	97.87	99.01	99.58
	Precision	63.87	88.58	72.01	98.9	88.16	65.44	90.33	82.06	99.40	89.44
	Recall	94.15	60.74	46.96	99.46	80.01	92.56	90.22	57.91	99.51	86.27
	F-score	76.09	79.73	56.78	99.18	83.85	76.67	90.27	67.90	99.46	87.83