Means for the formation of information fields and their inclusion in the unified information space of the region are analyzed, considering semantic connectivity and interoperability. The mathematical model for interaction of fields entering the information space is presented, conceptual apparatus and basic semantic relationships are described. The issues of representing subject field according to the principles of the system of systems proceeding from the division of analyzed and designed objects by the types of the relations connecting them and not by the types of elements are considered. Decomposition of these objects is proposed to be put into the basis of representation of subject field.

The results of automatic verification of applicant's curriculum vitae for compliance with a job vacancy using the algorithm based on natural language processing techniques are presented. The algorithm makes it possible to transform text information into the numerical features.

The method of continuous monitoring of unauthorized activity based on indirect parameters of the computing system as a means of ensuring its security is presented. The architecture of the information system of indirect monitoring is considered. The features of the implementation of its subsystems on the principle of inversion of control are studied. The possibility of integration with existing protection systems due to flexibility and scalability is determined, and further prospects for its development are analyzed.

The patterns of actual irreversible processes of aging of structural materials of equipment elements of structural and functional complex systems are theoretically substantiated. Mathematical models of various processes, which influence indicators of durability at submicroscopic, microscopic and macroscopic levels, are presented.

The aim of the study is to design the automation of the analysis of the results of power grid companies’ activity, to develop the system of assessment and criteria for indicators of financial condition, to emphasize the processes that can be automated. The design of an information system “The assessment of the financial condition of power grid companies”, which provides an efficient
activity of power grid companies due to in-time operational assessment and ability to accept valid management decisions in relation to the current state of the economics, is presented.

The concept of a medical information system for collaborative processing of phlebological data based on cloud technologies is described. The features of development and integration of the system of subsystems for collecting, preparing, accumulating of medical data (in particular, photographs and images of magnetic resonance imaging of the patients’ lower limbs) into the contour and their markings are considered. The development of the decision support subsystem for a phlebologist is presented. A comprehensive approach for intelligent processing of medical data in the formulation of the correct diagnosis in chronic venous insufficiency is proposed.

The aim of the study is to build a generator in a neural network basis and to create an algorithm for training a multilayer perceptron to solve the problem of pseudo-random values generation. Based on the results of the study on various distributions and sample sizes using computer modeling,
the high efficiency of the proposed architecture and training scheme for the algorithm is shown.

A formalized model of Cauchy type for knowledge assimilation based on the patterns of self-organization of educational process and its components is obtained in the framework of nonlinear
non-differential relationships with the possibility for dynamical change of model’s parameters, including the occurrence of chaos due to the bifurcation phenomena. New known mathematical
models, which consider both knowledge accumulation and dispersal in the educational process, for knowledge assimilation and assessment of remaining knowledge of students from universities are developed and adapted for educational digitalization.

The problem of selecting the optimum relationship between the accuracy and non-linearity is arising when constructing quasi-linear regression models. A two-criterial problem consisting of
maximizing the coefficient of determinant and simultaneous minimization of non-linearity criterion was formed to implement the previously developed criteria of non-linearity for area into the technology of arrangement of models “competition”. Two methods were proposed to solve the problem. The first method is based on the Pareto set, which makes it possible to visually select the most suitable alternative. The second method is more efficient from the calculating point of view as it eliminates
the highly non-linear variables. After that, the regression with the highest value of the determinant coefficient is being selected out of possible specifications. The proposed methods for solution
of two-criterial problem are demonstrated on the example of stimulating the operation of an evaporator.

The radiation of neutral particle that carries magnetic moment and executes non-inertial motion in the transverse magnetic field is analyzed. The power of radiation shows the asymmetric dependence on the horizontal spin components. The symmetry is restored for the ultra-relativistic and non-relativistic motions. At the same time, the first case shows the presence of radiation at either spin orientation whereas in the second case the radiation is absent when ς || H . In the case of uniform motion, the power expression obtained in the study coincides with the ones known from both the classical and the quantum theory.

The subject of the study is the efficiency of feature selection algorithms in relation to regression problems in the context of frequency of detecting false statistically significant dependencies.
The aim of the study is to build a suitable methodology and to test it on generated data, to test the hypothesis of frequency of occurrence of type II errors from distribution of dependent variable. In total, 7 methods of feature selection were studied in the work: Simulated Annealing, Select Difference,Hill-Climbing, Las Vegas, Sequential Forward Selection, Select Slope, Whale Optimization. Variables distributed according to 8 laws (Beta, Cauchy, exponential, Gamma, log-normal, normal, uniform, Weibull) were chosen as dependent variables. As a result of the study, it was found that the probability of using practical false-valued models is small using a rigorous approach in assessing the quality of models.