LiDAR: A Light-Weight Deep Learning-Based Malware Classifier for Edge Devices

Table 6

The evaluation results on the Raspberry Pi using the three classifiers.

Dataset	Classifier	CPU (%)	RAM (MB)	Classification time (seconds)	F1-score
SMS	CNNc	176.00%	264.21	3.31	0.94
	CNNg	169.00%	256.54	3.28	0.89
	LiDAR	84.70%	111.01	0.34	0.90
E-mail	CNNc	353.00%	2,029.13	130.73	0.99
	CNNg	345.80%	870.11	45.35	0.98
	LiDAR	167.50%	280.37	18.44	0.93
2019	CNNc	306.70%	582.38	7.51	0.94
	CNNg	294.10%	515.59	5.94	0.90
	LiDAR	189.60%	166.84	1.71	0.94
2020	CNNc	305.40%	638.52	8.15	0.99
	CNNg	303.40%	582.70	6.99	0.97
	LiDAR	172.80%	262.40	6.51	0.94