Table 1 Feature selection results for UNSW_NB15 and NSL-KDD datasets.
From: An effective method for anomaly detection in industrial Internet of Things using XGBoost and LSTM
Datasets | Threshold | Number of features |
|---|---|---|
UNSW_NB15 | 0.0005 | 50 |
0.001 | 39 | |
0.002 | 31 | |
0.003 | 23 | |
0.004 | 19 | |
0.005 | 17 | |
NSL-KDD | 0.0005 | 51 |
0.001 | 38 | |
0.002 | 28 | |
0.003 | 23 | |
0.004 | 17 | |
0.005 | 13 |
