APPLICATION OF MULTICRITERIA PREDICTION METHODS IN SYSTEMS FOR DETECTION OF NETWORK INTRUSIONS
Saint-Petersburg State University of Information Technologies, Mechanics and Optics; tutor
Abstract. Mathematical modeling of detection systems for network intrusion is considered. A method is proposed for network anomalies detection based on multicriteria prediction of security indices. Problem of computation of weighting coefficients accounting for time-dependent significance of prediction model is formulated and solved. The proposed approach is characterized by low enough computational complexity and does not depend on the type of criterion used for optimal prediction choice.
Keywords: intrusion detection, network anomalies detection, network security, multicriteria prediction.