MAE-CAD: An IP-Based Core Network Asset Discovery Technology Based on Multiple Autoencoders

Algorithm 1

The training and detection process of MAE-CAD.

	Input: The labeled dataset ; The unlabeled auxiliary dataset ; The dataset to be detected ; Hyper-parameters .
Output: The result dataset .
Step 0: Data preprocessing
Clean the data set to remove redundant and erroneous feature information.
Vectorize the data in the dataset , , and . And use MinMax in (1) for normalization
Step 1: Model construction based on the hyperparameters in .
Step 2: Pretraining
pretrained with dataset and .
Step 3: Fine-tuning
. in are pretrained.
trained with dataset .
Step 4: Detection
Result dataset
for in do
end for
return .