The paper is accepted by ECML-PKDD 2021.

Three public network intrusion datasets "KDD99", "NSLKDD", "UNSWNB15" and one synthetic dataset

Linux or macOS, CPU, Python 3, keras, numpy, pandas, scikit-learn.

git clone https://github.com/junzhuang-code/NEGMGAN.git
cd NEGMGAN/negmgan

For pip users, please type the command: pip install -r requirements.txt
For Conda users, you may create a new Conda environment using: conda env create -f environment.yml

1. gen_Data.py: Generate qualified dataset from raw data
2. preproc.py: Preprocessing the given datasets
3. utils.py: Utils modules
4. model_GMGAN.py: GM-GAN Models
5. model_Imeans.py: I-means Model
6. ablation_ExtractUCs.py: Ablation study -- Implement GM-GAN to extract UCs on streaming data.
7. ablation_DetectUCs.py: Ablation study -- Implement I-means algorithm to detect the number of new emerging classes on streaming data.
8. main.py: Implement NE-GM-GAN on one streaming batch data.

1. Generate datasets
   python [script_name] -DATA_NAME -DATA_DIR
   e.g., python gen_Data.py KDD99 ../data/

2. Ablation study

• Extract UCs: python ablation_ExtractUCs.py KDD99 1000
• Detect UCs: python ablation_DetectUCs.py Syn 2 10 3 200

3. Main function
   python [script_name] -DATA_NAME -NUM_EPOCHS -Z -WS
   e.g., python main.py Syn 1500 3 200