[Retracted] IoT-Oriented Distributed Intrusion Detection Methods Using Intelligent Classification Algorithms in Spark

Table 4

Comparative data of quintuple classification performance in recall, precision, and F1-score.

Evaluation indicator	Classification method	DoS	Probe	R2L	U2R	Normal
Recall (%)	Logistic regression	97.49	97.12	94.56	82.47	93.92
	Multilayer perceptron	99.73	98.61	92.85	50.51	95.86
	Decision tree	99.66	98.85	96.94	79.73	99.03
	Naive Bayes	83.07	83.11	78.06	64.86	81.73
Precision (%)	Logistic regression	98.86	94.85	78.69	70.54	97.43
	Multilayer perceptron	98.79	96	84.93	92.45	98.3
	Decision tree	99.15	98.8	92.45	76.62	99.03
	Naive Bayes	91.8	67.34	51.7	38.4	92.2
F1-score (%)	Logistic regression	98.17	96.05	85.9	74.77	95.65
	Multilayer perceptron	99.4	97.59	87.78	65.33	97.09
	Decision tree	99.63	98.83	94.4	78.15	98.58
	Naive Bayes	87.21	74.4	62.2	48.24	86.64