Tereykovskiy I. Neural models, methods and tools for evaluating Internet security settings-oriented information systems

The thesis deals with scientific and applied problems is to create an integrated development methodology widely-efficient neural evaluating the security settings of Internet-oriented information systems by theoretically informed choice characteristics, can efficiently identify new types of cyber attacks with limited computing resources variability and conditions of use. Based on the established methodology is based neural network evaluation system security settings, which in comparison with analogues can reduce the error classification, to verify the results and provide timely adaptation to the conditions of use and new types of cyber attacks. Using the proposed solutions developed means of identification of malware, spam, data leaks text and network cyber attacks.