The repository contains code refered to the work:

Giuseppina Andresini, Annalisa Appice, Luca De Rose, Donato Malerba

GAN Augmentation to Deal with Imbalance in Imaging-based Intrusion Detection

Please cite our work if you find it useful for your research and work.

 @article{ANDRESINI2021108,
title = {GAN augmentation to deal with imbalance in imaging-based intrusion detection},
journal = {Future Generation Computer Systems},
volume = {123},
pages = {108-127},
year = {2021},
issn = {0167-739X},
doi = {https://doi.org/10.1016/j.future.2021.04.017},
url = {https://www.sciencedirect.com/science/article/pii/S0167739X21001382},
author = {Giuseppina Andresini and Annalisa Appice and Luca {De Rose} and Donato Malerba}}

The code relies on the following python3.7+ libs. Packages needed are:

• Tensorflow 2.1.0
• Pandas 1.0.1
• Numpy 1.18.1
• Imbalanced-learn 0.7
• Hyperopt 0.2
• Keras 2.3.1
• Sklearn
• Pickle

The following DATASETS are trasformed to have a binary classification task (attacks or normal traffic). The datasets used are:

• CICIDS2017
• UNSW-NB15
• AAGM17
• KDD-CUP99

The repository contains the following scripts:

• main.py: script to execute MAGNETO
• train.py : script to execute the whole learning and testing task
• Cart2Pixel : script to create a mapping between examples into images
• ConvPixel : script that use a role to convert an array into images
• deep.py : script that contain two neural networks
• MinRect.py : script that calculate the minimum rectangle containing all the points in an array
• AGAN.py : script that create the ACGAN
• gan.py : script that train the ACGAN
• Generator.py : script to create examples from an ACGAN

Modify the following code in the main.py script to change the beaviour of MAGNETO

param = {"Max_A_Size": 10,  # Heigth and Weight of the images
         "Max_B_Size": 10, 
         "Dynamic_Size": False,  # search the minimum A and B to create 0 Collisions
         'Metod': 'tSNE',   # {tSNE, kpca, pca} to create the mapping between examples and images 
         "ValidRatio": 0.1, 
         "seed": 180,
         "dir": "dataset/dataset4/",  # path of dataset
         "Mode": "CNN2",  # Mode : CNN_Nature, CNN2
         "LoadFromPickle": False, # load dataset images from pickle
         "mutual_info": False,  # Mean or MI
         "hyper_opt_evals": 50, 
         "epoch": 200,
         "No_0_MI": False,  # True : remove 0 MI Features
         "autoencoder": False, # use autoencoder to reduce the number of features
         "enhanced_dataset": "gan"  # gan, smote, adasyn, ""None""
         }