The paper discusses the use of machine learning in relation to natural language processing (sentiment analysis, semantic proximity analysis) to build a recommendation system for the choice of perfumery products. The topic of the work is relevant in view of the growth of the range of manufactured perfumery products and the complexity of its choice by consumers and promotion by manufacturers. The proposed approaches are relevant for solving this problem due to the accumulated textual reviews and reviews of perfumery products on various websites, including online stores.

Keywords: machine learning, natural language, sentiment analysis, distributive semantics, word2vec, recommender systems

The results of the analysis of leakage currents as a function of time in samples of polycrystalline ferroelectric ceramics solid solutions of lead titanate-zirconate with sodium and bismuth impurities under the condition of applying a constant electric field are presented. We considered batches of samples that were stored in different conditions, which led to the formation of two batches - standard and defective. The samples of the standard and defective batches differed in values of the dielectric loss tangent by several times, while other electrophysical parameters practically did not differ. To measure the patterns of current fluctuations with time in the sample, a modernized Mertz scheme was used. The electric field was changed stepwise on the sample. Current fluctuations at a constant voltage across the sample were measured during the establishment of a constant leakage current and were considered as a time series, allowing only a statistical description. Changes in the form of the dependence of the leakage current on time are determined with an increase in the electric field strength on samples with different values of the loss tangent. The exponential decline is accompanied by random jumps in the magnitude of the leakage current. It is assumed that at a higher voltage, the state of the domain regions stabilizes, resulting in current leakage under a depolarizing field, since the domain system retains its more rigid configuration.

Keywords: mobile robots, polarization, depolarization, ferroelectrics, polarization current, remanent polarization, dielectric losses, electrical load, domain structure, electrical breakdown, mechanical failure

Parameters indicating the appearance of a glass-forming molten mixture in the melts of the salt system Ca(NO3)2 - KNO3 at a molar ratio of components of 3:2 are shown. The nonlinearity of the change in frequencies and line widths in the Raman scattering spectra of the anion NO3 was detected, which indicates a structural transformation in the molten mixture that correlates with a change in the viscosity of the mixture.

Keywords: melts, nitrates of alkali and alkaline earth metals, Raman spectra, binary salt systems

This work is devoted to modeling the absorption of microwave electromagnetic field energy in a device for milk pasteurization. Using CAD FEKO, a 3D model of the device was built and its operation was simulated. A study was made of the influence of the tilt angle of quartz tubes on the distribution of the electromagnetic field in the waveguide. The influence of the diameter and material of the tubes is also considered. The distributions of the electric field and the specific power absorbed per kg of dielectric in the waveguide are obtained. The accuracy of the results was assessed.

Keywords: pasteurization, microwave, heat treatment of dielectrics

The paper considers a model of security equipment reliability management using optimal control methods and data mining. An algorithm is given for finding a maintenance and repair management strategy that provides decision support in a wide range of system parameters changes. The possibility of automating the search for optimal solutions to the problem of increasing the reliability of security equipment is investigated on the example of the information system of analysis and predictive diagnostics of technical objects of the IS "Analyst", and the prospects for the use of neurofeedback in the tasks of increasing the reliability of complex technical systems are considered.

Keywords: security system, reliability of technical systems, optimal control, necessary conditions of optimality, neurofeedback

The widespread use of Educational Data Mining (EDM) allows the collection of data in the educational field for the analysis and adjustment of the educational process. Based on the analysis of the areas of training of construction specialties, a choice of job requests from the site was made hh.ru . A chain of nodes has been built in the Knime program for analyzing vacancies in construction specialties. Based on Knime queries, a semantic analysis of employers' requirements for construction specialties was carried out. The dependence of the result of the analysis of employers' requirements on the number of terms in the topic was considered. The preparation of the material was carried out to compare the data obtained from the work programs with the data extracted from the job search site hh.ru .

Keywords: analysis, data, work program, education, query, term, table, employer requirement

This article discusses the problems of constructing convolutional neural networks for determining road objects. The general relevance and formulation of the problem of determining road objects is presented. The rationale for the use of artificial neural networks for determining road objects has been formed. The Retinanet network architecture is used as the main architecture of an artificial neural network for determining road objects. The general concept of this architecture and the main subnets are visualized. Error functions for the main subnets of the Retina net network are described. The design description of algorithms for constructing data annotation for training an artificial neural network, as well as algorithms for constructing the neural network architecture of classification, regression and feature pyramid is given. The dynamics of changes in the general error function when determining road objects is determined. The result of training an artificial neural network is presented.

Keywords: convolutional neural networks, classification, regression, convolutional neural networks, deep learning, big data, mathematical modeling, computer science, RetinaNet architecture

The article describes a method for translating pictogram messages into text in Russian language based on machine learning. The text of the article contains the concept of alternative communication systems, a description of the training data preparing process, a description of the neural network architecture and the results of training.

Keywords: alternative communication systems, machine translation, machine learning, neural network, transformer architecture, Python, software

One of the key tasks in the development of the construction complex is to improve the quality and competitiveness of manufactured products and enterprises in general. The paper proposes a methodology for assessing the effectiveness of measures aimed at improving the competitiveness of construction companies and organizations, using the priority number of risk.

Keywords: 4P methodology, increasing competitiveness, performance evaluation, priority number of risk

This article discusses the aspects of collecting and processing textual data from social media and mass media in the context of assessing the social sentiments of the population of certain macro-regions. Text data is an important source of reflecting the sentiments of society and can be collected, pre-processed and analyzed using modern technical tools, considering the specifics of the hypotheses under study and the tasks set. This context is an important part of the system of comprehensive security of macro-regions in their cultural, social, economic, political aspects. The discussed approaches include the collection of relevant data, their processing and analysis using data mining tools.

Keywords: text analysis, media data, natural language processing, social sentiment assessment, data analysis

The question of creating an automated accompaniment is still an undisclosed part within the current automation of the musical field. The construction of accompaniment is used not only in the musical field, but also in related ones. Automatically generated accompaniment is used in audio and video studios for advertising, by people with and without musical education. In this paper, the existing methods for constructing automated accompaniment, audio file formats will be considered, and the developed algorithm and method for automatic accompaniment generation will be described.

Keywords: accompaniment, auto generated accompaniment, auto accompaniment, melody, MIDI

The article presents an approach to the development of a case-based decision support system for assessing the safety of a section of a city's transport system. To store the database of precedents, it is proposed to use an OWL ontology. The use of the ontological approach allows a more accurate interpretation of the subject area, and over time will expand and supplement information about the precedent. The precedent structure is defined. The ontology architecture is given: classes, properties and instances. An example of filling the ontology with precedents is given, using the example of the city of Volgograd. Measure of proximity to compare precedents is proposed. The procedure for adding new precedents to the knowledge base is considered. With the help of the developed decision support system, it is possible to assess the safety of both individual road sections and the entire transport infrastructure of the city.

Keywords: ontology, reasoning by precedents, transport, transport security

The article is devoted to the organization of data transmission from seismic stations via radio channel. The analysis of noise and noise immunity of phase-locked frequency systems, as well as the effect of noise on the synchronization of signals over time, is carried out.

Keywords: earthquake, data transmission, synchronization, encoding, frequency, noise, interference, refraction, time

The work is devoted to the analysis of the frequency characteristics of a piecewise homogeneous coaxial Bragg microwave structure with irregularities in the inner conductor. The article presents varieties of piecewise-homogeneous Bragg coaxial microwave structures, their options for practical application and implementation, as well as the possibility of their modernization, and also considers a recurrent method for calculating a piecewise-homogeneous transmission line. The possibility of using lumped capacitances in an equivalent circuit is considered and simulation is carried out in the electrodynamic simulation program with initial parameters.Based on the results of the work done, it was concluded that the effect of lumped capacitances included in the equivalent circuit of a piecewise homogeneous coaxial Bragg microwave structure significantly improves the frequency characteristics of a single-wave calculation in comparison with full-wave simulation. When approaching conditions favorable for the propagation of the first highest mode, it is necessary to include the effects of its propagation in the calculation.

Keywords: Bragg microwave structure, reactivity, permittivity, coaxial cable, microwave range

The article represents the results of the research of economic effectiveness of NIIOKR results. There are investigated the possibilities of using mathematic models for the evaluation of scientific production, the opportunities of its introduction into the economic practice.

Keywords: mathematic models, prognosis, economic effectiveness, economically valuable features, sorts

Temperature regimes of heating and cooling in three-dimensional printing are the most significant part of the technological solution in the production of products in any industry: from printing parts for aviation and rocket engineering to the construction of buildings or the manufacture of prosthetic human bones. The paper considers a simple and reliable method for obtaining calculated values of non-stationary temperatures and heat fluxes during layer-by-layer printing of products with various thermophysical properties and imperfect contact between layers, which does not require special software shells and large machine resources for calculations.

Keywords: non-stationary heat transfer, multilayer printing of products, additive manufacturing, 3D printing of products, heating and cooling of multilayer products, layer-by-layer deposition, optimization of the temperature regime of printing products

The paper considers a model for managing the reliability of a data transmission network used by an executive authority to organize access to state information resources. The process of constructing a computer model of the system for assessing the availability factor is described. The constructed computer model provides an opportunity to assess the reliability of the system in question in a wide range of parameters and can be used in the design, operation and modernization of the organization's information and telecommunications network to increase its reliability.

Keywords: information and telecommunication network, provider, routing, reliability, state information resource

In the process of conducting information activities, a large set of data accumulates, which reflects the specific features of the work performed. The stored information is not always in an orderly and understandable form, which makes it very difficult to work with it. This complicates the analysis, increases the processing time. Neural networks can solve this problem. Today, neural networks are widely used in many fields of activity, due to their application, for example, it becomes possible to analyze the market situation more thoroughly and make appropriate decisions that directly affect success. Thanks to the use of a neural network, it is possible to carry out a set of information in a convenient form for analysis. The article will provide a list of information about the self-organizing Kohonen map, concerning the principles of the neural network. The processing of test data with visualization of maps is considered.

Keywords: Kohonen self-organizing map, Kohonen networks, neural networks, cluster, processing, Self Organizing Map, SOM

The main directions of the energy strategy of railway transport are to improve the management structure of the railway energy complex, reduce the cost of electric energy and reduce the cost of its acquisition. The initial information for planning optimal operating modes in the management of the energy complex is provided by the forecast of electricity consumption. According to the rules of functioning of retail markets, consumers are required to accurately plan the volume of electricity consumption. If the power consumption deviates by more than 5% of the planned volume, the company incurs additional costs. To make an accurate forecast, it is necessary to analyze the source data – the archive of electricity consumption. At the initial stage of data analysis, the problem of omissions is revealed. If there are gaps in the data, the process of forecasting electricity consumption can be difficult, and sometimes impossible. The most rational solution is to fill in the gaps using modern methods of information processing. This will allow you to clearly present the data structure, calculate the necessary values and interpret the results of the analysis.

Keywords: power consumption, time series, data gaps, recovery of missing values, forecasting of power consumption, train traction, railway transport, neural networks

The construction of a regression model of errors in the approximation of the current curve from factors affecting this error is described. The following factors are selected as influencing factors: the number of coefficients of the Bessel-Fourier decomposition and the number of points on which the original function is constructed. Experimental data were obtained as a result of modeling electrical processes occurring in a pulsed magnetization reversal system.

Keywords: permanent magnet, regression model, Bessel-Fourier decomposition, approximation, momentum, factors, function, curve, momentum, error

The task of planning the trajectory of a mobile robot is considered in the following formulation: it is necessary to choose the optimal trajectory of the robot's movement over hilly terrain from one point to another among the many available paths. The developed algorithm for planning the optimal trajectory of movement of mobile objects allows minimizing energy costs when moving over hilly terrain. The optimal trajectory is determined from the condition of minimum of the functional of the rolling friction forces work. To find the optimal trajectory on which the friction forces work is minimal, is used the dynamic programming method. The proposed algorithm for planning the optimal trajectory can be used to reduce energy consumption in the process of moving over hilly terrain not only autonomous robotic systems, but also other vehicles.

Keywords: planning algorithms, rolling friction force operation, dynamic programming, optimization according to the selected criterion

The article examines the issue of calculating the depth of compression stresses when using a spherical indenter in order to increase the fatigue strength of steel parts. When the spherical indenter is pressed, elastic deformation of the surface first occurs, and then plastic one. The residual plastic deformation is expressed in the size of the print. The physicomechanical properties of the processed material are not taken into account, the process of pressing a spherical indenter into a steel sample in the ANSYS program is modeled by the final element analysis (FEA).A bilinear deformation diagram was fed, which is set in the program by the modulus of elasticity, yield strength and Poisson's ratio. At the next stage, a finite element grid was generated by using the Hex Dominant Method, which divided the model into elements in the form of hexahedra of various sizes. In the contact area, the size of the finite elements was 0.2 mm. The results of calculations of analytical and numerical methods for spherical indenter diameters of 4, 6, 8, 10 mm and various degrees of plastic deformation to determine the depth of compression stresses are presented, compared with known data, the error is calculated. A dependence is derived for a preliminary assessment of the occurrence of residual stresses depending on the diameter of the well after the indentation of the spherical indenter.

Keywords: spherical indenter, bilinear deformation diagram, plastic deformation, indentation depth, compression stress