A.S. Shaburov - Ph.D. (Eng.), Associate Professor, Department of Automation and Telemechanics, Perm National Research Polytechnic University. E-mail: shans@at.pstu.ru

The improvement of data processing systems complicate considerably the decisions about information defense. The appliance of neural network for solving the defense tasks in informational systems is the long-term area for researches. The ability to educate students on neural networks is used in modelling of systems which are difficult to forecast. The models, based on neural networks, obtain the ability to adapt to dynamic conditions and function fast. This is especially relevant while information defense systems are working. Currently, multilayer architectures of neural networks as well as multilayer perceptron are used. The training of multilayer perceptron is based on backward propagation of errors. The choice of the neural network structure is implemented in accordance with peculiarities and difficulty of tasks, resolving by this network. The neural network can be the multilayer perceptron, which implements linear and non-linear shrink of input data in the buried layer. Such neural network is called a return network. The recognition of computer attacks in the process of informational system functioning can be presented, basing on settings analysis of system work according to stated rules. The defense attribute space is formed, decomposed and systematized relying on precious experience and knowledge about the running of informational system itself. As a result, defense attribute space is specified on the base of a posteriori information. The recognition of a computer attack is probable, basing on the integration of return neural systems and multilayer perceptron, which are connected sequentially. The training type of neural systems is defined by the method of setting enrollment. The process of training of a neural system can be implemented in accordance with the mathematic problem definition. The solution of a such problem is predisposed by numeral optimization methods.

 

1. Rashevskijj R.B., SHaburov A.S. O primenenii nejjronnykh setejj dlja obespechenija bezopasnosti informacionno-upravljajushhikh sistem kriticheski vazhnykh obektov // Nejjrokompjutery: razrabotka primenenie. 2014. № 12. S. 31-35.
2. Rashevskijj R.B., SHaburov A.S. Prakticheskoe primenenie nejjronnykh setejj dlja zashhity informacionno-upravljajushhikh sistem kriticheski vazhnykh obektov ot DDOS-atak // Nejjrokompjutery: razrabotka, primenenie. 2015. № 10. S. 16-20.
3. Rashevskijj R.B., SHaburov A.S. Aktualnye ugrozy bezopasnosti informacionno-upravljajushhikh sistem kriticheski vazhnykh obektov // Nejjrokompjutery: razrabotka primenenie. 2013. № 11. S. 96-100.
4. Markov A.S., Fadin A.A. Organizacionno-tekhnicheskie problemy zashhity ot celevykh vredonosnykh programm tipa Stuxnet // Voprosy kiberbezopasnosti. 2013. № 1(1). S. 28-36.
5. Rutkovskaja D., Pilinskijj M., Rutkovskijj L. Nejjronnye seti, geneticheskijj algoritmy i nechetkie sistemy. M.: Gorjachaja linija - Telekom. 2013. 384 s.
6. Kornienko A.A., Sljusarenko I.M. Sistemy i metody obnaruzhenija vtorzhenijj: sovremennoe sostojanie i napravlenie sovershenstvovanija // Internet statja. 2009.
7. Mironova A.A., SHaburov A.S. O trebovanijakh k razrabotke modeli kompjuternykh atak na telekommunikacionnye sistemy // Innovacionnye tekhnologii: teorija, instrumenty, praktika (INNOTECH 2015): materialy III Mezhdunar. Internet-konf. mol. uchen., aspirantov, studentov, 01 - 30 nojab. 2015. Perm: Izd-vo PNIPU. 2016.
8. Klimov S.M., Sychev M.P., Astrakhov A.V. Protivodejjstvie kompjuternym atakam. Metodicheskie osnovy // EHlektronnoe uchebnoe izdanie. M.: MGTU imeni N.EH. Baumana. 2013. 108 s.