Trust in Intrusion Detection Systems: An Investigation of Performance Analysis for Machine Learning and Deep Learning Models

Table 6

The KDD dataset separated 70% training set and 30% testing set.

Attack type	Training set (70%)	Testing set (30%)
Dos	249609	106896
Normal	54826	23590
Probe	2672	1125
R2l	784	341
U2r	25	12
Sum	307916	131964