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

Table 5

Class imbalance ratios of Train A, Train B, Validation, and Test dataset.

Classes	#Train A	Ratio (%)	#Train B	Ratio (%)	#Validation	Ratio (%)	#Test	Ratio (%)
Normal	97,278	24.52%	48,639	10.13%	48,639	24.52%	60,593	26.15%
U2R	52	0.01%	26,026	5.17%	26	0.01%	39	0.01%
R2L	1,126	0.23%	254,476	92.70%	563	0.23%	5,993	2.09%
DoS	391,458	381.68%	195,729	58.73%	195,729	381.68%	223,298	323.61%
Probe	4,107	0.84%	4,108	0.78%	2,053	0.84%	2,377	0.82%