The paper investigates the numerical simulation of wave processes in multilayer thin films, which is relevant for understanding their physical properties and optimization for various applications. An integrated mathematical model has been developed that combines Maxwell’s equations, mechanical vibrations and thermal conductivity, taking into account the interaction of physical fields in structures with defects. Adaptive algorithms have been proposed for automatic mesh refinement depending on local gradients of physical parameters, which allows to increase the accuracy of modeling in critical zones. A hybrid approach to calculations using CPU and GPU has been implemented, which ensures efficient use of resources for large-scale problems. Software with a modular architecture has been developed that allows integrating numerical methods, optimization and visualization of results in real time. Experimental validation has confirmed the high accuracy and reliability of the model. The results obtained contribute to a deeper understanding of physical processes in thin films and are the basis for the creation of highly efficient multilayer structures in industrial and scientific applications.
The solution of the scientific and practical problem of determining the effect of factor influence on the result of the work of the information measuring system for the technological process of manufacturing processed cheese is considered through the use of a factor influence model that takes into account the simultaneous effect of five factors and their cross-interactions on the control indicator. The task of the study is to implement for use a simplified cross-classification model that makes it possible to estimate the amount of expected information about the levels of the control parameter when taking into account the levels of both influencing factors and their mutual interactions. An electrical schematic diagram of the control system has been developed and its practical implementation has been carried out, thanks to which statistical data on the main parameters of the technological process have been obtained. Conclusions have been drawn about the possibility of further use of the proposed cross-classification model for various information measuring systems regardless of their purpose.
The publication is devoted to analyzing electrocardiogram signals of artificial origin of realistic form with the possibility of controlling the duration, sampling frequency, noise level and pulse rate using Ingrid Daubechies wavelets. The synthesized signals were investigated using discrete wavelet analysis to study the influence of these parameters on the approximation and detail coefficients. The priority influence of noise on the detail coefficients and the dependence of the number of signal peaks on the given parameters were established. The article uses for the first time the method of packet discrete wavelet filtering of detail coefficients and approximation coefficients. This allowed to provide a high degree of signal restoration to the original form. Similar studies were conducted for continuous wavelet transformation with the generation of wavelet scalogram images, which provide additional diagnostically significant information. The results obtained in the form of an algorithm are promising for use in analyzing signals from radar systems. The developed model for generating realistic-shaped signals is more efficient and exceeds the average accuracy (96.2 %) compared to analogues (88.03 %). The effectiveness of the developed method is fully confirmed by the correlation matrix of functions of discrete spectra of arti-ficial ECG signals.
This article explores the issue of the application of Kolmogorov–Arnold Networks (KAN) for biological age estimation using a dataset of 344 male patients. The dataset includes biomarkers related to bone health and body composition. To enhance model performance, data preprocessing techniques such as polynomial interpolation for missing values and standardization were applied. Pearson and Spearman correlation analyses identified the most relevant biomarkers. Machine learning models were evaluated, along with neural networks and KANs. Experimental results demonstrate that KANs outperform traditional machine learning models and classical neural networks on small datasets. The optimal KAN architecture achieved a correlation coefficient of 0.93, a mean squared error of 18.81, and a mean absolute error of 2.8, surpassing the best-performing conventional models. These findings highlight the potential of KANs as a robust alternative for biological age estimation in resource-limited settings.
The important task has been solved in this scientific research related to specific development and verification of the fundamental suitability of the method for automatic synthesis of Petri nets based on the functioning of an artificial neural network. This allows to automate the generating process of logical control algorithms. It has been proposed to divide the artificial neural network into separate local networks for the subsequent stepwise training. Such stepwise training is necessary to form the logical control algorithm and to synthesize the appropriate Petri net. The proposed principles can be applied in the early stages of training for an artificial neural network while synthesizing a Petri net. It has been conducted the experiments linked to the training of local neural networks and to implement the formation of an algorithm for logical control of the laboratory machine for processing chemical solutions. As a result of the experiment, it has been determined the fundamental suitability for the stepwise training method of neural networks for generating Petri nets and the automated formation of logical control algorithms.
This study introduces a method for determining the best locations for pressure sensors in water supply networks and for assessing network conditions using artificial intelligence techniques. The goal is to identify the network nodes that would provide the most important information for detecting water leaks and evaluating the overall network status. The selection of sensor locations was based on data sets of pressure changes caused by various leak scenarios generated by EPANET simulations. Genetic algorithms were used to rank candidate nodes and determine the optimal number of sensor locations. The next step involved assessing the network state using the ANFIS neuro-fuzzy network and the Mamdani neuro-fuzzy logical inference algorithm. These algorithms were implemented in the Google Colab environment and tested on a section of the water supply network in Kyiv, Ukraine.
This paper substantiates the relevance of addressing the problem of ensuring the quality of change descriptions in source code files within version control systems. To filter commit messages, machine learning methods are employed, including neural networks of various architectures. The use of neural networks is justified by the need to identify descriptions that accurately reflect the intent of the changes. A comparative analysis of word embedding methods (Word2Vec, FastText, and GloVe) was conducted, along with their application in binary classifiers such as MLP and RNN for filtering code changes. The models were trained on a dataset of change descriptions collected via the GitHub REST API. Model performance was evaluated using Accuracy and F1-score metrics. The effectiveness of the Google Colab environment for prototyping machine learning models was also confirmed.
Improving the quality of IT services that support accessibility standards is a way to increase the inclusiveness level of digital content. Taking into account the individual requests of users with disabilities in the technical characteristics of software products ensures their digital accessibility. The goal of the article is to develop a comprehensive approach to the application of Software Quality Function Deployment (SQFD) technology to improve the quality of support for IT service accessibility standards. The SQFD method was implemented by using the developed ArtiHoQ software solution, which made it possible to establish connections between consumer quality indicators and technical characteristics of the service, as well as to determine the target values of technical characteristics. The proposed measures to improve technical characteristics were implemented on the example of the uRemediate IT service to ensure accessibility for the users with disabilities, which made it possible to come close to the target values for all key metrics, and significantly increase the comprehensive quality indicator.
Recent advances in large language models (LLMs) enable the automated discovery of semantic structures and emerging signals within text streams, offering an opportunity to redesign foresight workflows into continuous, data-driven systems. This study aims to develop and validate an automated framework for extracting, structuring, and comparing semantic ontologies using LLMs. The paralyzed approach was used for data mining from social media platforms and filtering non-domain data. The key semantic elements, goals and hypernyms corresponded, were extracted using multiple LLM configurations, with a consensus mechanism to provide semantic reliability and minimize hallucination. The extracted elements were embedded in a high-dimensional vector space, clustered iteratively using cosine similarity, and merged hierarchically. Convergence process and structural stability were analyzed using the elbow criterion and similarity metrics. The Proposed approach provides a cost-efficient alternative to traditional expert-based foresight analysis. By integrating LLM-driven semantic extraction with quantitative clustering, it enables the identification of emerging trends, weak signals, and long-term thematic structures. The results highlight the potential of LLM-based semantic modeling as a foundation for automated foresight systems.
A method for solving the traveling salesman problem is proposed, utilizing dynamic programming to determine the shortest duration route, considering the fuzzy representation of travel time between individual points. Approaches for the approximation of fuzzy values, arithmetic operations, and methods for ordering fuzzy numbers are presented. The problem formulation with trapezoidal fuzzy numbers is considered. A representation form of such fuzzy numbers based on a Gaussian-like approach is proposed. Methods of the state space tree and dynamic programming are employed. The proposed technique is illustrated with an example.
Text data clustering is a key component of unstructured text message analysis. To utilize these methods, text data must be converted into vector representations, i.e., word embeddings must be performed. This paper presents a modification of the HDBSCAN* clustering algorithm using custom distance metrics from the Minkowski family (L1, L2, L∞) and parameters specifically tailored for clustering unstructured text data. A major contribution is a novel evaluation metric based on the relative point density of identified clusters and surrounding noise formations (“clouds”). Beyond assessing overall clustering quality, this metric highlights problematic dense accumulations within the noise that require additional manual analysis. Experimental evaluation on the “20 Newsgroups” dataset demonstrated that clustering quality is independent of the α parameter but highly sensitive to the distance metric, with L∞ yielding the best results. The nomic-embedding-v1 model significantly outperformed gte-v1.5 in both the silhouette score and the proposed relative density metric.
This research presents the mathematical derivation of curvature tensors from the Schwarzschild solution of the gravitational field, along with the derivation of gravitational waves as the second derivative of the field. The elimination of the event horizon is the core contribution of this paper, building on Paul Dirac’s approach in his 1975 book, General Theory of Relativity. This paper introduceslinear and non-linear models for the spacetime distortion function. The fit of the tensor equation for gravitational waves is tested using geometrical information from the LIGO-Virgo observation of WG200115, employing econometric techniques with indicators such as R-squared for the robustness of the given data and the fit of the model to the data distribution, as well as the Durbin–Watson statistic for time-series predictability. The results show a good fit of the derived mathematical model to the observed geometry. Different models for simulating spacetime distortion exhibit varying degrees of fit to the observed geometry, with a less distorted model providing a better explanation for the observed gravitational waves. The model’s fit decreases when considering the rotation of the source object, whether it be a neutron star or a black hole. Further modeling effort is needed to accurately represent the NSBH merger process and its role in the formation of gravitational waves.
The realization of novel scenario involving transitions between different types of chaotic attractors is investigated for the Rössler system. Characteristic features indicative of the presence of generalized intermittency scenario in this system are identified. The properties of “chaos–chaos” transitions following the generalized intermittency scenario are analyzed in detail based on phase-parametric characteristics, Lyapunov characteristic exponents, phase portraits, and Poincaré sections.
This paper proposes an examination of effective methods and tools for ensuring software quality. The scope of this topic includes current issues related to software quality assurance within the context of analyzing methods and tools used in practice to develop high-quality software. During the modeling process, a new comprehensive model for software quality assurance has been developed, combining modular testing, integration testing, and continuous integration methods. The advantage of this development is its enhanced adaptability to addressing key challenges in software quality assurance. Based on the developed model, strategies and approaches are proposed to improve configuration management processes and identify vulnerabilities in software systems.
A multi-criteria optimization mathematical model of credit scoring is proposed. The model is derived using a nonlinear trade-off scheme to solve multi-criteria optimization problems, allowing for the construction of a Pareto-optimal solution. The proposed approach forms an integrated assessment of a borrower’s creditworthiness based on a structured set of indicators that reflect the financial, credit, and social profile of clients. The model is designed for use in intelligent CRM and ERP systems operating on Big Data and does not rely on labeled training samples, making it applicable to unsupervised learning tasks. It can also serve as a foundational layer for further deep-learning analysis. Methodological steps for implementing the model, from indicator normalization to final decision-making, are described. A technological implementation demonstrates the model’s effectiveness in automated loan decisions and fraud detection.
Prediction of the toxicity of chemical compounds is one of the most important steps in drug design. The use of phenolic compounds is a promising component in the pharmaceutical industry with many possible applications. The paper fo-cuses on the application of a probabilistic neural network for classifying 232 phenols based on their mechanisms of toxic action. The Kruskal–Wallis test was also used to assess the influence of molecular descriptors on the reliable classification of phenolic compounds based on the mechanisms of their toxic action. It is shown that for the correct training of a probabilistic neural network and effective prediction of the mechanisms of toxic action of phenols, it is suf-ficient to use only 5 molecular descriptors.
The problem of applying generalized linear models to the analysis of actuarial risks in the context of premium charges to clients was considered. The Monte-Carlo method for Markov chains was applied. Two situations were considered for the computational experiment. For the first one, insurance indicators and the target variable were randomly assigned due to the problem of public data access. To create three datasets, charges were generated from normal, gamma, and Pareto distributions with dynamic variance, and noise was added to stimulate a non-stationary process. In the second situation, actual actuarial data from the Singa-pore Actuarial Society was used. Generalized Linear Models with normal dis-tribution and logarithmic link function, an exponential distribution and loga-rithmic link function, and Laplace distribution with identity link function were constructed. Based on the model-fitting quality metrics, conclusions were drawn about their structure.
This article provides a brief overview of a set of the most common basic object detection neural network models. Today, the need for automating surveillance and observation processes remains a growing trend. Moreover, one of the key tasks of such processes is usually the detection of an object of interest for further analysis. Previously, many basic object detection algorithms and approaches have been proposed; however, most of them typically have limitations in terms of their applicability. In most cases, these limitations arise due to the nature of the observed environment or because the detection approaches rely on specific object characteristics, such as color or basic shapes only. To address these problems, a new approach for object detection has been developed using neural networks. This paper presents the basis and central aspects of the most common neural network object detection models. The experiment has demonstrated the features, advantages, and disadvantages of the studied methods in the application case of lab animal detection during their behavioral study. Considering this, conclusions and recommendations for their usage cases were made.
This article examines evaluation metrics for the results of super-resolution image generation in solving the SISR task. The study comprises two experiments: the implementation of custom network architectures for SRGAN, VDSR, and SRCNN, and fine-tuning of pre-trained SRGAN, VDSR, and SRCNN models. An algorithm for assessing the quality of models and deep learning methods for generating super-resolution images is suggested. The VDSR model performed best in terms of pixel, structural, and perceptual metrics, as well as training time and visual confirmation by a human, highlighting that residual learning is more effective than recursive learning under the conditions of the two conducted experiments. Threshold values for practically acceptable and high-quality results were determined through visual analysis of many generated images and their corresponding quality metrics, including those reported by other researchers.
The efficiency of a multi-position system depends on the realization of its struc-ture — how many elements it includes, where they are located, and how the envi-ronment and terrain influence its operation. The paper is dedicated to data pro-cessing in a multi-position surveillance system as an additional option, leverag-ing the in-between big data from the system’s elements. A sufficient number of numerical data generated by the multi-position system and its elements–sensors–allows the use of statistical methods and models from machine learning or deep learning. The ontology for quality estimation of the multi-position sys-tem, depending on its configurations, is proposed. The results of the distribu-tions of detected events are presented in graphical forms that allow statistical evaluation of the distributed data. Our findings allow us to ensure the efficiency of a multi-position system in an unpredictable, variable environment by recon-figuring it when it offers better capabilities.
