Detecting Malware with an Ensemble Method Based on Deep Neural Network

Table 5

The malware detection experiment results for different methods.

Models	Accuracy (%)	AUC	TPR (%)	FPR (%)	EER (%)	Training time (h)	Detection time (ms)
-gram	93.21	0.9864	89.22	0.1	3.94	4.18	2304
LSTM	99.13	0.9999	98.69	0.1	0.54	1.34	13.62
CNN	98.14	0.9989	96.92	0.1	1.55	1.46	15.97
LSTM + CNN + MF (MalNet)	99.88	0.9999	99.14	0.1	0.37	2.91	30.33