Research Article
Empirical Evaluation of Noise Influence on Supervised Machine Learning Algorithms Using Intrusion Detection Datasets
Pseudocode 4
The fourth set of experiments (noise filtering then injection).
(1) | Load the training portion of the intrusion dataset | (2) | Convert the label values from numeric to nominal (only for UNSW-NB15) | (3) | Conduct noise filtering by removing the outliers and the extreme values | (4) | Conduct noise injection by manipulating the labels of the training instances (5, 10, 20, and 30%) | (5) | Train the ML algorithm on the noisy training dataset | (6) | Run the ML algorithm on the testing dataset | (7) | Document the results | (8) | Revert to step 4 until each of the ML algorithms is trained and tested with the different levels of noise |
|