MalwareClassifier

Requirements

• this project uses jupyter notebooks, for tutorial on usage see jupyter-notebook
• requirements.txt
  • contains modules needed to execute the code which are not included in the standard python distribution (apart from the jupyter notebook)
  • to install run: pip install --requirement requirements.txt

Project structure

• folder Scripts contains:
  • TestDataExtraction and TrainDataExtraction subfolders.
    • Scripts within these folders are responsible for extraction of training and test data
  • TrainFeaturesScaling
    • Scales all training samples into the range [0, 255]
  • SingleFeatureProcessing
    • Contains overview of each feature together with a classifier based on each feature
  • WholeFEatureSetClassification
    • Contains LeNet-5 and random forest classifier for the whole feature set
  • utilities
    • contains helper classes and functions
• folder Data.zip:
  • contains only the final version of training and test features in feature_merged_scaled and Data/test_data/test_features_merged_scaled
  • the whole dataset is not provided as it is several GB in size, but it can be found in the Ember dataset
  • upon downloading the following version: ember_dataset_2018_2.tar.bz2 both training and test data can be extracted using corresponding extract_to_csv.ipynb jupyter notebook