IoT-IE: An Information-Entropy-Based Approach to Traffic Anomaly Detection in Internet of Things

Table 1

Detection performance for individual devices.

Algorithm	AE			TP			NC			CU			LX
	Acc (%)	TPR (%)	FPR (%)	Acc (%)	TPR (%)	FPR (%)	Acc (%)	TPR (%)	FPR (%)	Acc (%)	TPR (%)	FPR (%)	Acc (%)	TPR (%)	FPR (%)
KNN	98.5	97.0	0.0	98.9	98.7	0.4	96.8	95.3	1.6	94.6	91.0	1.7	99.6	99.2	0.0
DT	98.5	100.0	3.1	98.2	99.1	4.5	95.2	94.6	4.2	93.3	91.7	5.0	99.4	98.9	0.0
RF	99.5	99.0	0.0	99.4	99.4	0.4	96.5	95.1	2.1	94.2	92.0	3.6	99.4	100.0	1.2
SVM	100.0	100.0	0.0	99.1	99.0	0.4	97.5	96.5	1.6	94.3	91.1	2.5	99.4	99.2	0.4
NB	99.5	99.0	0.0	98.9	99.0	1.2	85.9	88.8	17.0	86.1	89.3	53.2	99.2	98.9	0.4