Identification of Biometric Images by Machine Learning

Nazarkevych, Mariya; Voznyi, Yaroslav; Hrytsyk, Volodymyr; Klyujnyk, Ivanna; Havrysh, Bohdana; Lotoshynska, Nataliia

doi:10.1109/elit53502.2021.9501064

Cited by 6 publications

(1 citation statement)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the nomenclature is not limited to the three mentioned groups. For example, in the SEIR model 15 , group E (exposedinfected, not contagious) was added to the three mentioned groups. All these models are instances of the family of compartmental epidemiological models [16][17][18][19][20] .…”

mentioning

confidence: 99%

Stochastic forecasting of variable small data as a basis for analyzing an early stage of a cyber epidemic

Kovtun,

Grochla,

Kharchenko

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Security Information and Event Management (SIEM) technologies play an important role in the architecture of modern cyber protection tools. One of the main scenarios for the use of SIEM is the detection of attacks on protected information infrastructure. Consorting that ISO 27001, NIST SP 800-61, and NIST SP 800-83 standards objectively do not keep up with the evolution of cyber threats, research aimed at forecasting the development of cyber epidemics is relevant. The article proposes a stochastic concept of describing variable small data on the Shannon entropy basis. The core of the concept is the description of small data by linear differential equations with stochastic characteristic parameters. The practical value of the proposed concept is embodied in the method of forecasting the development of a cyber epidemic at an early stage (in conditions of a lack of empirical information). In the context of the research object, the stochastic characteristic parameters of the model are the generation rate, the death rate, and the independent coefficient of variability of the measurement of the initial parameter of the research object. Analytical expressions for estimating the probability distribution densities of these characteristic parameters are proposed. It is assumed that these stochastic parameters of the model are imposed on the intervals, which allows for manipulation of the nature and type of the corresponding functions of the probability distribution densities. The task of finding optimal functions of the probability distribution densities of the characteristic parameters of the model with maximum entropy is formulated. The proposed method allows for generating sets of trajectories of values of characteristic parameters with optimal functions of the probability distribution densities. The example demonstrates both the flexibility and reliability of the proposed concept and method in comparison with the concepts of forecasting numerical series implemented in the base of Matlab functions.

show abstract

mentioning

confidence: 99%