Experts at EURid registry of the European domain .EU together with researchers from the University of Leuven (Belgium) developed a method to identify domain names that could be used for illegal purposes. The method was presented on the RAID (Research in Attacks, Intrusions and Defenses) cyber security conference. It is based on the unique algorithms of recognition and machine learning techniques.
Official EURid website informs that at the moment the methodology allows to discover domain names that will be used for illegal purposes in the future (such as spam mailings, spreading malware, creating points of sale for counterfeit goods, etc.) in 80% of the cases. The share of false positives doesn't exceed 1%. This month already EURid plans to test software based of this methodology. It will flag suspicious newly registered names for a more detailed study of them. The testing will last for 6 months. In case of success, on the next stage software will be used to prevent activation of suspicious names.
