Machine-Learning Approach to Optimize SMOTE Ratio in Class Imbalance Dataset for Intrusion Detection

Table 17

Confusion matrix of RNN-LSTM.

		Predicted class
		Normal	U2R	R2L	DoS	Probe
Actual class	Normal	59,528	92	657	86	230
	U2R	4	24	10	0	1
	R2L	4,316	23	1640	12	2
	DoS	20	0	2	223,169	107
	Probe	4	0	0	6	2,367