Convolutional Neural Networks for DGA Detection

Just a set of R scripts for evaluating a simple 1D Conv neural networks for detecting DGA.

Usage:

The main script loads train and test files in RData format and performs a 5-fold Cross validation. Default dataset is loaded in csv format from datastes/argencon.csv

$ Rscript ./evaluate_dga_classifier.R

By default the script saves the resulting train and test files in the format used by the CNN. The following execution of the script will use these files. If required they can be regenerated using the parameter --generate

Results files

The script saves two files in csv format in the results/ directory. The first contains info about the cross validation results and the second information per malware family.