Adaptive DDoS Attack Detection Method Based on Multiple-Kernel Learning

Table 3

Comparison results of four algorithms for amplifying the normal flow.

		The value of random multiplier
		1.0–1.5	1.5–2.0	2.0–2.5	2.5–3.0	3.0–3.5	3.5–4.0	4.0–4.5	4.5–5.0	5.0–5.5
ADADM method	DR (%)	78.93	78.93	78.93	78.93	78.93	78.93	78.93	78.93	78.93
	FR (%)	0.01	0.01	0.01	0.47	1.42	1.90	4.74	6.64	10.43
	ER (%)	12.02	12.02	12.02	12.22	12.63	12.83	14.05	14.87	16.50
SimpleMKL method	DR (%)	77.14	77.14	77.14	77.14	77.14	77.14	77.14	77.14	77.14
	FR (%)	0.01	0.01	0.47	1.42	4.27	6.64	10.90	17.06	21.80
	ER (%)	13.04	13.04	13.24	13.65	14.87	15.89	17.72	20.37	22.40
SVM method	DR (%)	77.86	77.86	77.86	77.86	77.86	77.86	77.86	77.86	77.86
	FR (%)	0.01	0.01	0.95	1.42	3.79	8.06	11.85	18.01	22.75
	ER (%)	12.63	12.63	13.04	13.24	14.26	16.09	17.72	20.37	22.40
Nezhad et al.’s [16] method	DR (%)	97.49	97.13	97.13	97.13	96.77	96.77	96.77	96.77	96.77
	FR (%)	65.24	65.24	66.19	66.67	66.67	66.67	67.14	67.14	67.62
	ER (%)	29.47	29.67	30.08	30.28	30.49	30.49	30.69	30.69	30.90