A Deep Random Forest Model on Spark for Network Intrusion Detection

Table 7

The experiment results on the four given datasets.

Dataset	Algorithm	Accuracy	Precision	Recall
NSL-KDD	PRF	0.93224	0.94648	0.93013
	DSSVM	0.94137	0.95763	0.95065
	A-DNN	0.98952	0.98101	0.97474
	FS-DPRF	0.99132	0.99213	0.98733
UNSW-NB15	PRF	0.85321	0.93752	0.87308
	DSSVM	0.81927	0.92931	0.86054
	A-DNN	0.94219	0.95740	0.93829
	FS-DPRF	0.97763	0.98363	0.96401
CICIDS2017	PRF	0.86532	0.85655	0.86227
	DSSVM	0.85290	0.84739	0.84075
	A-DNN	0.96527	0.96028	0.94272
	FS-DPRF	0.97430	0.95330	0.97931
CICIDS2018	PRF	0.89876	0.91010	0.90733
	DSSVM	0.87852	0.89110	0.88031
	A-DNN	0.94480	0.94754	0.93015
	FS-DPRF	0.96750	0.96031	0.95424