Hybridization of computational intelligence methods for attack detection in computer networks

摘要: Abstract The paper is devoted to identification and classification of network traffic connections by various hybridization schemes with the goal efficient attack detection. For this purpose combination different methods computational intelligence used, namely neural networks, immune systems, neuro-fuzzy classifiers support vector machines. To increase speed processing input vectors it proposed apply method principal components. A distinctive feature advantage approach suggested a multi-level analysis traffic, providing possibility detect attacks signature based technique combining set adaptive detectors on methods. describes software tool that built basis mechanisms. Computational experiments were carried out serve as evidence their effectiveness in detection both known unknown attacks.