Adaptive DDoS Attack Detection Method Based on Multiple-Kernel Learning

Table 2

Comparison results of four algorithms for narrowing the attack flow.

		The value of the random multiplier
		0.1–0.2	0.2–0.3	0.3–0.4	0.4–0.5	0.5–0.6	0.6–0.7	0.7–0.8	0.8–0.9	0.9–1.0
ADADM method	DR (%)	78.21	78.21	78.21	78.57	78.57	78.57	78.57	78.57	78.57
	FR (%)	10.99	1.42	0.01	0.01	0.01	0.01	0.01	0.01	0.01
	ER (%)	17.15	13.04	12.42	12.22	12.22	12.22	12.22	12.22	12.22
SimpleMKL method	DR (%)	75.71	76.43	76.43	76.43	76.43	76.43	76.43	76.43	76.43
	FR (%)	22.75	4.74	1.42	0.01	0.01	0.01	0.01	0.01	0.01
	ER (%)	23.63	15.48	14.05	13.44	13.44	13.44	13.44	13.44	13.44
SVM method	DR (%)	76.07	77.14	77.50	77.86	77.50	77.86	77.86	77.50	77.86
	FR(%)	22.75	4.74	1.42	0.47	0.01	0.01	0.01	0.01	0.01
	ER(%)	23.42	15.07	13.44	12.83	12.83	12.63	12.63	12.83	12.63
Nezhad et al.’s [16] method	DR(%)	97.13	97.85	98.57	98.21	98.57	98.57	98.21	98.21	97.85
	FR(%)	74.29	74.29	74.29	74.29	74.29	74.29	74.29	74.29	74.29
	ER (%)	33.54	33.13	32.72	32.92	32.72	32.72	32.92	32.92	33.13