Using a Subtractive Center Behavioral Model to Detect Malware

Table 10

Comparison of n-gram and the proposed model (400 malware and 300 benign).

Model	Classifier	DR (%)	FPR (%)	Acc. (%)
4-gram	J48	91.4	9.1	91
	LMT	97.7	2.4	97.4
	RF	85.1	18.8	85
	SLR	94.6	6.3	94.5
	SMO	92	9.6	92.1
	KNN	87	16.2	87.3
	BN	—	—	—
	NB	86.7	16.4	87
Proposed model	J48	99.5	0.7	99.4
	LMT	98.6	1.5	98.4
	RF	96.1	4.9	96
	SLR	98.5	4	97.2
	SMO	97.4	2.4	97.3
	KNN	87.4	13.6	87.7
	BN	86.6	12.8	86.5
	NB	75.8	20	75.5