Komar M. Intelligent Information Technology for Attacks Detection and Classification in Information Telecommunication Networks

The thesis is devoted to topical issues of the development of new methods for the detection and classification of attacks on information telecommunication networks (ITN) by using artificial neural networks and artificial immune systems, and implementation of intelligent information technology that is based on mentioned above, and the assessment of its reliability. The method designing the neural network detector attacks on ITN is improved, and it's based on neural network of vector quantization which uses 80% of the neural elements in the hidden layer that meet the type of attack, and the last - to the normal connection which is characterized by a small number of the training set. The method designing a comprehensive classifier is improved for the hierarchical classification of network attacks. This method is based on multi-channel neural network detectors, which is uniting the principal component analysis, consolidation and elimination of conflicts between the neural network detectors, each of which is trained for a certain type of attack, which allowed reducing the dimension of the analyzed information and classifying the network attacks. There is developed a combined method that is based on the integration of neural network detectors in an artificial immune system, this allowing them to adapt to unknown attacks with the help of cloning and mutation operations. On the basis of proposed methods was developed further intelligent information technology for detecting and classifying attacks on ITN by using the basic principles of the immune system in order to create the best population of detectors. Developed technology is characterized by the generating the detectors set per each type of network attack.