Table 6 Original and SMOTE balanced sets.

From: Comparative evaluation of data imbalance addressing techniques for CNN-based insider threat detection

Dataset instances

Original dataset

Training set

Test set

SMOTE balanced training set

Malicious instances

346

272

74

1,846,778

Non-Malicious instances

2,308,467

1,846,778

461,689

1,846,778

Total

2,308,813

1,847,050

461,763

3,693,556