This is my project about Ransome Detection using GAN, the experiment was implemented by cuckoo.
- read report files from specific location (set in cuckoo) and extract features from them
- features engineering on choosing features with highest weights
- generate matrix based on the features(occurences of features)
- experiments with black attack for the GAN model
- experiments based on different parameters, number of hidden layers, number of nodes
- experiments based on ensemble classifiers
- experiments based on ensemble neural networks
- experiments on both ransomware dataset and generous malware
- collect dataset:
- please put the generated reports seperately under dataset folder
- run the col_reports inside the processing (specify your own folder)
- generate raw features:
- run the the feature_gen to generated seperate features
- get the total unique feature list:
- run the feature_dict to get features list
- get the database based on feature occurrences
- run the feature_fin to get dataset.csv
- classifier analysis (in classifiers):
- analyse and filter the generated datasets and classifier algorithms
- model experiment:
- test the models with filter features
For the ransomware and cuckoo configuration, please check the link cuckoo-download-instructions