Research Article
Deep Encrypted Traffic Detection: An Anomaly Detection Framework for Encryption Traffic Based on Parallel Automatic Feature Extraction
Algorithm 2
Parallel automatic feature extraction algorithm.
| Input: | | Preprocessed Dataset | | Hyperparameter = {Minbatch, LR, Epoch, Bias} | | Output: | | Feature Set | (1) | for each one in do | (2) | for epoch in (1, Epoch) do | (3) | for each batch of Minbatch data do | (4) | for in batch do | (5) | for each in do | (6) | Calculate the output of the jth SAE network | (7) | Process Sigmoid | (8) | Compute the cost between output and input | (9) | Update the weight and bias | (10) | end for | (11) | end for | (12) | end for | (13) | end for | (14) | Connect the exported feature |
|