Malicious software is a security problem that has been around for many years. The topic that is currently, being extensively investigated is botnet detection. Botnets are global networks of compromised computer, that a botmaster can use to for example run Distributed Denial of Service attacks campaign.
Time and time again, researcher have to create good data sets for training and testing the machine learning algorithms.
The goal of this project is to create a framework of a system which is publicly available, and that enables easy comparison between various machine learning based botnet detection methods, where each detection method is tested with the same data sets for training and testing.
This also requires extensive knowledge about best practices in capturing, labelling and merging data sets into training and evaluation sets.