Fuzzy preference approach for computer network attack detection
Miodrag Manić, Bogdan M. Wilamowski · 2002
Irrepressible growth of complex interconnectedness of information systems besides its obvious benefits unfortunately brought up the questions of their vulnerability. Practically universal access to computers has enabled hackers and would-be terrorists to attack information systems and critical infrastructures worldwide. A fuzzy preference relation, based on a fuzzy satisfaction function is applied to a comparison of attack signatures. Fuzzy signatures (their gamma resolution sets) are combined by fuzzy operators. Therefore, a qualitative, fuzzy decision system is achieved. The proposed method provided smoother results that obtained by traditional methods. Experiments demonstrated that the final outcome depended on correct determination of fuzzy values out of signature attacks, also on adequate choice of fuzzy set operator.