K.V. Sazonov1, M.V. Tatarka2

1 Military University of Radio Electronics (Cherepovets, Russia)
2 18 Federal State Unitary Enterprise Research Institute of the Ministry of Defense of the Russian Federation (Moscow, Russia)
 

The active development of information and telecommunication networks, as well as the constant increase in the number of objects participating in mutual information exchange (users of Internet resources), imposes special requirements for information security, one of the areas of support, which is the formation of dynamic profiles and constant monitoring of the nature of actions users. Profiling means the creation of a dynamic image of an object that takes into account a multitude of technical parameters of various categories, both informational and technical. The creation of a set of dynamic profiles will make it possible to rank Internet objects according to the criteria of importance within the framework of the tasks of ensuring information security for the purpose of their subsequent monitoring.

Target – Improving the effectiveness of monitoring Internet objects in order to early detect and prevent illegal activities in the information and telecommunication space.

The recognition algorithm in the form of an abstract system is considered, the main information components of the object identification system are described. Within the framework of the task of identifying Internet objects, the theoretical principles for constructing pattern recognition systems are generalized. A data structure model is formulated, mathematically described and practically implemented, taking into account the main aspects of the pattern recognition problem. Thus, this model, taking into account the possibility of iterative changes in parameters, allows you to analyze machine learning models based on observing the dynamics of the quality metric of interest or a given group of metrics. At the same time, a statistical analysis of real data obtained on existing samples of the parameters of objects of infotelecommunication networks allows you to vary the parameters of the model within the limits necessary for analysis. To analyze the probability of recognizing infrastructure objects depending on the received parameters, their type, volume, number of features, etc., as well as for a deeper understanding of the feature descriptions generated by the model, let's consider a number of examples.

The results obtained are the basis for the analytical description and software implementation of a multi-parameter model for the formation of attribute descriptions of objects, the analysis of which, within the framework of solving the problem of identifying or classifying Internet objects, makes it possible to reasonably (scientifically) make decisions on the necessary requirements for the set of collected parameters of network objects. These requirements allow you to determine the main characteristics of the data set necessary to solve the problem with the necessary quality metrics.

Sazonov K.V., Tatarka M.V. Multiparametric model for the formation of feature descriptions of objects for studying the properties of classification algorithms. Science Intensive Technologies. 2022. V. 23. № 7. P. 50−70. DOI: https://doi.org/10.18127/j19998465-202207-05 (in Russian)

1. Vapnik V. Chervonenkis A. Teoriya raspoznavaniya obrazov (statisticheskie problemy obucheniya). M.: Nauka. 1974. 416 s. (in Russian).
2. Chervonenkis A. Komp'yuternyj analiz dannyh: Lekcii shkoly analiza dannyh YAndeks. M.: Yandeks. 2009. 259 s. (in Russian).
3. Flah P. Mashinnoe obuchenie. Nauka i iskusstvo postroeniya algoritmov, kotorye izvlekayut znaniya iz dannyh. M.: DMK Press. 2015. 400 s. (in Russian).
4. Tatarka M.V. Identifikaciya abonentskih terminalov informacionno-telekommunikacionnyh setej̆ na osnove modeli formirovaniya izobrazheniĭ v sovremennyh komp'yuternyh sistemah. Trudy SPIIRAN. 2020. T. 19. № 2. S. 446–474 (in Russian).
5. Tatarka M. V., Cazonov K.V., Shuvaev F.L. Metod strukturnogo statisticheskogo raspoznavaniya rezhimov raboty informacionno-vychislitel'nyh setej̆. Informaciya i kosmos. 2019. № 1. S. 40–46 (in Russian).
6. Vlasov A.I., Zaharov E.R., Zaharova V.O. Nejrosetevaya sistema obnaruzheniya i nejtralizacii udalennogo nesankcionirovan-nogo vmeshatel'stva v komponenty Interneta veshchej. Nejrokomp'yutery: razrabotka, primenenie. 2021. T. 23. № 1. S. 63–80.