Application of High-Dimensional Outlier Mining Based on the Maximum Frequent Pattern Factor in Intrusion Detection

Table 2

The result of two mining algorithms.

Sample set (sample size)	Threshold value	Accuracy (%)	False positive rate (%)
			Normal	DoS	Probing	R2L	U2R
Normal + DoS (63000)	58500	88	1	Null	2	0	2
	59100	92	1	Null	2	1	4
	60000	90	1	Null	2	1	3
Normal + Probing (62000)	58000	88	2	3	Null	1	10
	59000	95	2	4	Null	1	10
	61000	92	2	2	Null	1	10
Normal + R2L (60900)	59000	87	3	0	2	Null	31
	59600	95	6	0	2	Null	33
	60000	93	7	0	2	Null	35
Normal + U2R (60052)	59000	77	3	0	0	48	Null
	59500	87	3	0	2	50	Null
	60500	83	4	1	2	50	Null