Computer systems and information technologies

ANALYSIS OF FRESCHET AND HAUSDORF METRICS AND THEIR MODIFICATIONS FOR IMAGE COMPARISON

Mykola BEREZKYI — Thu, 25 Sep 2025 00:00:00 +0300

This paper provides a comprehensive analysis of classical and modern metric approaches used for quantitative evaluation of image similarity, including the Fréchet and Hausdorff distances as well as their modifications – the Gromov-Fréchet and Gromov-Hausdorff metrics. The relevance of this research is determined by the wide use of image comparison methods in computer vision systems, where they form the basis for segmentation, classification, and object detection in various application domains, particularly in medicine. Images are represented as polygons, which unifies computational procedures and simplifies the formal description of distance measurement algorithms.

The properties of the considered metrics were compared, and computational experiments demonstrated that the Fréchet distance effectively reflects the similarity of polygon contours, while the Hausdorff distance is more suitable for comparing inner regions. The Gromov-based modifications provide minimal distances and more flexible results when dealing with objects of complex structure. Algorithmic solutions for each metric are described, with an emphasis on their computational complexity and possible practical applications. Special attention is given to isometric transformations, which reduce matching errors.

The results were validated through experiments implemented in Java with the OpenCV library, proving the adequacy and efficiency of the proposed approaches. The practical value of the research lies in the potential integration of the obtained results into automated medical diagnostic systems for the analysis of histological, cytological, and immunohistochemical images. The proposed algorithms may serve as a basis for developing effective segmentation and classification methods for biomedical data.

CONSTRUCTION OF A FORMAL MODEL OF THE EXCURSION SELECTION PROCESS FOR STUDENTS: A CASE STUDY OF THE “PUSH” SCHOOL IN KHARKIV IN THE FORM OF A LOGICAL AFP-NETWORK

Iryna VECHIRSKA, Alina SOKOLOVA, Natalia VALENDA — Thu, 25 Sep 2025 00:00:00 +0300

The article is devoted to the study of the finite predicate algebra toolkit for constructing a network model aimed at formalizing the process of selecting educational excursions for students of general secondary education institutions, in particular for the “Push” School in Kharkiv. The objective of this work is to develop a corresponding mathematical model in the form of a logical AFP-network. The organization of excursions constitutes an important component of the educational process, as it enables the integration of learning, recreation, social interaction, and career guidance for students. In the context of modern challenges—such as ensuring safety, aligning with educational objectives, and optimizing costs—there is a clear need for the development of effective methods for planning and selecting excursions, which determines the practical relevance of the research problem.

The application of predicate algebra in this domain has made it possible to formalize the excursion selection process while accounting for a wide range of parameters, including cost, duration, safety level, student categories, interests, educational objectives, accommodation conditions, and others. The methodology for constructing logical networks provides a systematic analysis of the subject area and enables the creation of a mathematical model in the form of a complex polyadic relation through the composition of corresponding binary relations derived from the subject area analysis, thereby optimizing the processing of input knowledge.

The mathematical model of the excursion selection process is represented as a polyadic predicate dependent on a set of variables, each associated with a specific domain of definition. Knowledge processing is carried out in the network nodes through conjunction and disjunction operations, with knowledge represented as predicates corresponding to subsets of the respective domains. Each binary predicate is described in the form of bipartite graphs and the corresponding formulas. The conjunction of all constructed binary predicates forms a logical network that enables iterative processing of information until a stable result is achieved.

The scientific novelty lies in the developed model, which facilitates the automation of the excursion selection process while accounting for the individual characteristics of students and the resources of the school. The results of this study represent an effective tool for optimizing educational programs.

INFORMATION AND LASER SYSTEM FOR ASSESSING HARMFUL MAN-MADE EMISSIONS

Liubomyr SIKORA, Nataliia LYSA, Olga FEDEVYCH — Thu, 25 Sep 2025 00:00:00 +0300

At the current stage of development in the energy, chemical, mechanical engineering, and printing industries, there is extensive use of environmentally hazardous resources such as coal, oil, natural gas, paints and coatings, and polymer materials. The high production intensity driven by market demands leads to increased consumption of energy and raw materials, which in turn causes a rise in concentrations of dust, toxic gases, and harmful liquid emissions into the atmosphere and water systems. This contributes to growing environmental pollution, the real-time assessment of which is often impeded by the limitations of conventional data collection methods.

In response, new approaches have been substantiated for the development of sensors capable of measuring levels of air contamination, particularly dust and toxic substances. These innovations are based on novel physical principles, enabling the creation of laser-based pollution concentrators, sensors utilizing the optogalvanic effect, and integrated systems that combine ion-selective sensors (such as OCM 5M) with measurement platforms. Such technologies enhance the overall effectiveness of environmental monitoring and safety systems.

A comprehensive solution to this issue involves the establishment of global environmental monitoring networks that rely on information and intelligent technologies, as well as the development of next-generation sensor models. Environmental monitoring has remained a critical issue for over a century, as industrial progress — from railroads and textile production to thermal power and petrochemical complexes — has brought not only prosperity but also significant environmental degradation. The impact has been especially pronounced with the advancement of nuclear energy and jet aviation, and further intensified by the consequences of the First and Second World Wars. More recently, the war in Ukraine has introduced a unique set of ecological threats, including explosions, damage to energy infrastructure, and the destruction of oil terminals.

RESULTS OF APPLICATION OF INFORMATION TECHNOLOGY FOR PROCESSING AND ANALYSING ELECTROCARDIOGRAM SIGNALS TAKING INTO ACCOUNT THEIR MORPHOLOGICAL AND RHYTHMIC CHARACTERISTICS

Andriy SVERSTIUK, Lyubomyr MOSIY — Thu, 25 Sep 2025 00:00:00 +0300

This paper presents an information technology for comprehensive processing and analysis of electrocardiographic signals considering their morphological and rhythmic characteristics based on a cyclic random process mathematical model for improved cardiac pathology detection. The proposed information technology utilizes a cyclic random process (CRP) model that naturally accounts for the quasi-periodic structure of ECG signals with variable rhythm. The system implements sequential and parallel processing through interconnected functional blocks: preprocessing (baseline drift removal, noise filtering), automatic segmentation for cycle boundary detection, rhythm function formation using cubic spline interpolation, cyclic signal transformation ensuring equal sample counts per cycle, and separate statistical processing of morphological and rhythmic features. Experimental verification was performed on real ECG signals from patients diagnosed with atrial fibrillation and atrial flutter.

The technology successfully segmented ECG signals and quantified both amplitude and temporal variability. For atrial fibrillation, the rhythm function demonstrated significant variability (400-800 ms range) with sharp transitions between adjacent values. For atrial flutter, the rhythm function showed greater stability (150-230 ms range) with smoother fluctuations. Statistical analysis revealed distinct patterns of morphological variability, with dispersion values reaching 0.004 mV² for atrial fibrillation and 0.025 mV² for atrial flutter. The key innovation lies in the simultaneous yet separate analysis of morphological and rhythmic characteristics through rhythm function incorporation, enabling comprehensive assessment of both amplitude variability and beat-to-beat dynamics within a unified CRP framework. The developed technology enables automated differentiation between various cardiac pathologies through independent classification of rhythmic and morphological abnormalities, supporting clinical decision-making in cardiac diagnostics and real-time monitoring applications.

INFORMATION SYSTEM FOR ADAPTIVE TRANSPORTATION PLANNING WITH CONSIDERATION OF ROAD TRAFFIC VARIABILITY

Hlib ISHCHENKO, Oleksandr SHEVCHUK — Thu, 25 Sep 2025 00:00:00 +0300

This paper presents an information system for adaptive transportation planning that integrates the classical linear-programming transportation problem with a time-dependent cost model approximated by a combination of normal probability density functions. The proposed mathematical model implements discretization of the 24-hour interval into equidistant time steps, which enables correct accounting for diurnal variations in road conditions while preserving high computational efficiency. A unified operational algorithm was developed based on the classical Simplex method, and key methods for constructing an initial feasible plan for the transportation problem were implemented to allow comparative analysis of performance and accuracy in dynamic conditions.

The outcome of the study includes an intuitive web interface implemented with React, using react-vis for charting, Leaflet for interactive maps and OSRM for routing, together with a server module written in Go that employs the gonum/lp library for solving linear-programming problems. The proposed architecture provides fast interaction between client and server modules, high scalability and straightforward cross-platform deployment.

Experimental validation confirmed the correctness of the model both in cases with static cost coefficients and in the enhanced time-dependent transportation formulation. In particular, the system supports automated temporal analysis of solutions and identification of cost-optimal departure times. In the intercity scenario dynamic optimization yielded up to 7.2 % savings relative to the worst static scheduling alternative, while in the urban scenario accounting for time-dependent costs produced savings up to 47.8 % for evening departures compared to the typical morning peak — consistent with observed urban traffic patterns.

A comparative analysis with leading commercial transport management systems demonstrated that, despite its streamlined architecture, the proposed system delivers the required level of flexibility and adaptivity while markedly reducing implementation and maintenance costs. Consequently, it constitutes an accessible and transparent tool for educational institutions, research activities and local logistics projects in small and medium-sized enterprises.

ADVANCED METHODS OF APPLYING CODING SYSTEMS IN THE DESIGN OF DIGITAL COMPONENTS FOR CYBER-PHYSICAL SYSTEMS

Natalia VOZNA, Volodymyr HRYHA, Lesya MYCHUDA, Lidiya SHTAIYER — Thu, 25 Sep 2025 00:00:00 +0300

This study aims to develop highly efficient digital components for cyber-physical systems capable of self-recovery, high-speed data processing, and reliable operation in real-time environments. The paper proposes new approaches to information encoding, particularly for RGB color image representation, and the modeling of neuro-like structures using the code systems of Krestenson, Rademacher, and Haar. The methodology is based on the mathematical foundations of the residue number system (RNS), modular arithmetic, and structural analysis of digital components. Hybrid models of formal neurons, perceptrons with delay lines, and wavelet neurons are employed to solve signal classification tasks. A model of signal self-recovery in a neural bundle is developed, taking into account failures of individual elements and inhibitory effects. As a result, a fault-tolerant mechanism for information transmission in bioneural structures is implemented, along with algorithms for encoding RGB pixels in the R-C and H-C code systems. These methods ensure unambiguous decoding and allow adaptive encoding in the presence of data loss or corruption. The scientific novelty lies in the integration of biological fault-tolerance principles with digital encoding methods based on RNS, providing adaptive signal recovery without the need for complete decoding. The practical significance of the research is in the potential application of the results to digital vision devices, sensor platforms, embedded systems, and high-performance processors for intelligent computing in cyber-physical environments.

IMPLEMENTATION OF INFORMATION TECHNOLOGY FOR THE MODEL OF PREDICTION POTENTIAL CARS COLISION

Oleksandr BYZKROVNYI, Kyrylo SMELYAKOV — Thu, 25 Sep 2025 00:00:00 +0300

This research focuses on developing an information technology solution capable of delivering real-time notifications about potential car accidents to Internet of Things (IoT) devices installed in vehicles. Since vehicles are mobile and frequently change location, there is a need to have stable and timely notifications. To address this, the proposed solution leverages cloud technologies, which offer scalability, low latency, and high availability for real-time data transmission.

The main objective is to design and implement architecture that enables effective real-time messaging between cloud services and moving IoT devices. The system is tested across various simulated driving scenarios to evaluate its performance and reliability. A key aspect of the methodology involves measuring the time interval between when a message is sent and when it is received by the IoT device, followed by assessing whether the driver has adequate time to respond appropriately.

The experimental setup is based on test cases created using Python, the BeamNG.tech simulation environment, and AWS IoT Core as the cloud service provider. The results demonstrate that the proposed technology can reliably handle real-time messaging in dynamic conditions. As a conclusion, the research confirms the potential of the developed solution for real-life applications in connected vehicle safety systems.

ANALYSIS OF ELECTRICITY CONSUMPTION USING THE COMPONENT METHOD OF PERIODICALLY CORRELATED RANDOM PROCESSES

Andrii VOLOSHCHUK, Halyna OSUKHIVSKA, Mykola KHVOSTIVSKYI, Andriy SVERSTIUK — Thu, 25 Sep 2025 00:00:00 +0300

Contemporary energy systems, considering the diverse challenges emerging within energy infrastructure, require advanced analytical methodologies for electricity consumption forecasting. Traditional statistical approaches prove insufficient for modeling dynamic multi-scale temporal structures of electricity consumption signals aimed at predicting electrical loads in residential households.

This research presents a comprehensive approach to electricity consumption analysis utilizing the mathematical framework of periodically correlated random processes (PCRP), specifically employing the component method. The mathematical foundation of the methodology consists in representing electricity consumption signals as PCRP models with decomposition into constituent elements: deterministic trend components, periodic components of cyclical variations, and stochastic components of random deviations. Component analysis enables the identification of latent consumption patterns through decomposition of periodic characteristics. Therefore, the proposed method allows for the elimination of limitations inherent in traditional stationary models.

Empirical validation was conducted using a comprehensive dataset of residential electricity consumption spanning the period from July to August 2025. Experimental data demonstrated pronounced repetitive characteristics with systematic daily periodicity, confirming the theoretical premise regarding daily component dominance. Three-dimensional visualization of results revealed complex interaction dynamics between different frequency components of electrical load signals. Spectral analysis exhibited characteristic distribution with maxima for low-frequency components corresponding to daily harmonics.

The obtained results can be utilized for residential electrical load forecasting and enable both short-term and medium-term energy consumption predictions. This is significant not only for forecasting residential electrical loads, but also for optimizing electrical energy resources and managing intelligent networks.

SYNTHESIS OF RECURSIVE DEVICES FOR VERTICAL-GROUP CALCULATION OF BASIC MULTI-OPERAND NEUROOPERATIONS

Ivan TSMOTS, Oleh BEREZSKY , Taras MAMCHUR — Thu, 25 Sep 2025 00:00:00 +0300

The operational basis of artificial neural networks has been determined, comprising groups of the following neurooperations: preprocessing, processing, and computation of transfer functions. A set of basic multi-operand neurooperations was selected for hardware implementation, including: finding maximum and minimum values in a one-dimensional data array, calculation of the sum of squared differences, scalar product calculation. The methods of vertical-group computation of basic multi-operand neurooperations (finding for maximum and minimum values in a one-dimensional array, calculation of the sum of squared differences, and scalar product calculation) have been improved. Using the selection of number of bits for operands group for single-cycle processing, these methods enable synchronization of data arrival time with calculation time and ensure high hardware utilization efficiency during the hardware implementation. It's proposed a recursive devices design for vertical-group computation of basic multi-operand neurooperations based on an integrated approach. This approach leverages the capabilities of modern element base, incorporates vertical methods, algorithms, and recursive device structures for implementing basic neurooperations and considers the requirements of specific applications. The principles for designing recursive devices for vertical-group calculation of basic multi-operand neurooperations have been chosen. These include: the use of a basis of elementary arithmetic operations and a multi-operand approach; modularity; pipelining and spatial parallelism; homogeneity and regularity of the structure; synchronization between data arrival time and neurooperation calculation time; specialization and adaptation of structure to specific application requirements. A format converter has been developed to transform a flow of serial input data from a one-dimensional array into a parallel-serial data output by group of bits. Basic structures have been developed. They represent calculation algorithms in terms of hardware and serve as the foundation for synthesizing of recursive devices for vertical-group calculation of basic multi-operand neurooperations with specified parameters. The method for synthesis of recursive devices for calculation of basic multi-operand neurooperations with vertical-group data processing has been improved. Through the use of mechanisms for synchronizing calculation time with data arrival time, this method provides the selection of structure which performs real-time data processing and with high hardware utilization efficiency. It has been demonstrated that the use of the improved vertical-group methods, designed basic structures of devices for finding maximum and minimum numbers in one-dimensional arrays, calculation of the sum of squared differences and scalar product, as well as the improved synthesis method, enables real-time mode and the implementation of devices for calculation of basic multi-operand neurooperations with vertical-group data processing with high hardware utilization efficiency.

INFORMATION SYSTEM FOR SERVICE PORTFOLIO FORMATION FOR INFOCOMMUNICATION PROVIDERS

Thu, 25 Sep 2025 00:00:00 +0300

This paper addresses the challenge of service portfolios formation for infocommunication service providers. A comprehensive analysis of existing approaches enabled the formulation of the problem within the framework of contemporary business models and underscored the necessity of developing new effective methods. These methods must reflect the specifics of interaction between IT companies, service providers, and end users, as well as the inherent characteristics of services and their delivery environments. The paper substantiates the choice of methodological foundations underlying the proposed system for service portfolio formation. The formal problem is categorized as a nonlinear, multicriteria Boolean programming task. Growing demands for alignment between service packages and user needs, coupled with the resource constraints of modern IT infrastructures and the complexity of inter-service dependencies, highlight the need to solve large-scale optimization problems. To address this, a hybrid method is proposed, combining problem decomposition into subproblems, the application of metaheuristic techniques for their resolution, and heuristic procedures for integrating partial solutions. The paper presents experimental results demonstrating the effectiveness of the proposed approach. Additionally, it outlines system-level design solutions that support the development and implementation of the information system for service portfolio formation.

ANALYSIS OF OPTIMIZER AND HYPERPARAMETER INFLUENCE ON YOLO IN THERMAL LANDMINE DETECTION

Natalia MELNYKOVA, Anna VECHIRSKA — Thu, 25 Sep 2025 00:00:00 +0300

This paper investigates the impact of optimizer choice and hyperparameter tuning on the performance of the YOLO deep learning model for landmine detection in thermal images. The aim of this work is to study the effect of different optimizers and parameter configurations on model accuracy and training stability. The object of the study is the process of detecting landmines in thermal imagery using deep neural networks.

A dataset of thermal landmine images annotated in YOLO format was used for training. The experiments were conducted with the YOLOv11n architecture initialized with pre-trained weights. The varied parameters included the optimizer (SGD or Adam), learning rate, and batch size. Each model was trained for 50 epochs, and performance was evaluated using mAP, precision, and recall metrics.

The study provides a comparative analysis of the influence of Adam and SGD optimizers on the accuracy and stability of YOLO when trained on a limited dataset of thermal landmine images. The results suggest that, given appropriate configuration, SGD is capable of achieving performance competitive with adaptive methods, despite their popularity. The experiments also confirm the feasibility of achieving high detection accuracy even with a relatively small dataset.

All configurations achieved high mAP values. The Adam optimizer enabled a faster initial reduction in loss functions, whereas SGD provided smoother and more stable training dynamics. The highest precision and recall were obtained in the experiment with SGD at a learning rate of 0.01 and batch size of 64, making this configuration the most promising for further research.

The findings on optimizer and hyperparameter selection can be applied to improve the efficiency of automated thermal image analysis systems based on unmanned aerial vehicles, contributing to safer and faster detection of explosive hazards.

The METHOD FOR ASSESSING THE QUALITY OF FINGERPRINT COMPARISON BASED ON CONVOLUTION OF BOOLEAN METRICS

Yurii POHULIAIEV, Kirill SMELYAKOV — Thu, 25 Sep 2025 00:00:00 +0300

The proposed approach is designed to determine the degree of similarity between fingerprints. It is based on the comparison of key features of papillary patterns. This method can be used to compare both a pair of fingerprints and one fingerprint with a group of others. Within the framework of this approach, basic metrics are used that reflect the presence of matches between the compared fingerprints. Based on these metrics, the following characteristics are calculated: the percentage of matching features and the weighted average score, taking into account the significance of each match. These parameters are used to establish the belonging of a fingerprint to a certain group. To control the results, threshold values are used, established on the basis of the Bayesian classifier and the Neyman-Pearson lemma, taking into account the differences in the distribution of scores. Testing conducted using the FVC2000 database showed that the accuracy of comparing individual fingerprints is 93.81%. The method is implemented in Python using the Pandas and NumPy libraries. This allows it to be integrated into automated fingerprint identification systems (AFIS). Resistance to moderate distortions and interpretability of parameters indicate its practical applicability, in particular, in the field of forensics. One of the limitations of the method is its susceptibility to strong fingerprint distortions. Further development plans include the introduction of adaptive thresholds and integration with deep learning methods to improve the process of feature extraction on fingerprints. This development is aimed at increasing the accuracy and reliability of biometric identification, which is extremely important for reducing the likelihood of errors, especially in the context of forensic tasks.

EFFICIENCY ANALYSIS OF FINANCIAL TIME SERIES FORECASTING MODELS UNDER MARKET TURBULENCE CONDITIONS

Oleh PASTUKH, Yuriy PETROV — Thu, 25 Sep 2025 00:00:00 +0300

This paper presents a comparative analysis of financial time series forecasting models' effectiveness under market turbulence conditions. The study focuses on evaluating the adaptability of statistical ARIMA model and recurrent LSTM neural network across different prediction horizons during periods of high market volatility. Daily OHLC data from five major technology companies (Google, Apple, Amazon, Meta, Oracle) for the period 2020-2025 was analyzed, with particular emphasis on the turbulent April-June 2025 period. Three model architectures were implemented: ARIMA(2,1,0), LSTM Bidirectional Autoencoder (100 units), and simple LSTM (20 units). Testing was conducted on 5, 15, and 30-day forecasting horizons using MAPE, RMSE, and MAE metrics. Additionally, residual analysis through autocorrelation function examination was applied to validate model quality. Results demonstrate that ARIMA excels in short-term forecasts (5 days) with MAPE ≤ 0.06, but its effectiveness diminishes on medium-term horizons due to inability to adapt to market turbulence. Simple LSTM (20 units) achieved optimal balance between accuracy and stability, outperforming ARIMA by 30.75% on medium and long-term forecasts. Complex LSTM Autoencoder proved least effective due to overfitting on market noise. The scientific novelty lies in comprehensive analysis of model adaptability to extreme market turbulence using residual analysis as additional validation method. It was proven that simpler LSTM architectures outperform complex ones under high volatility conditions. The practical significance includes optimization of algorithmic trading strategies and risk management systems during market instability periods, particularly valuable for financial institutions and investment funds.

LEGAL AND ETHICAL BASES FOR CREATING REPRESENTATIVE DATASETS TO DETECTING MANIFESTATIONS OF CYBERBULLYING IN TEXT CONTENT

Olena SOBKO, Archil CHOCHIA — Thu, 25 Sep 2025 00:00:00 +0300

The article is devoted to developing the method for creating of representative text data datasets for detecting manifestations of cyberbullying in text content, considering ethical and legal principles. The primary focus is ensuring fair and equal representation of different demographic groups in text samples, which is critical for creating non-discriminatory and socially responsible artificial intelligence models. Emphasis is placed on compliance with key ethical principles – preventing harm, avoiding bias, and ensuring representativeness – and provisions of international law, particularly the General Data Protection Regulation. Proposed method for creating of representative text data datasets for detecting manifestations of cyberbullying in text content, taking into account ethical principles, which includes the following stages: preliminary processing of text data, analysis of distributions according to ethical aspects (age, gender, religion etc.), and representative adjustment through multi-criteria optimization. Machine learning models are trained on prepared balanced samples using appropriate reference datasets to classify text samples according to ethical criteria. The comparison is based on official demographic data for Ukraine, which ensures the reliability of the assessment of deviations.

As a result of applying the developed method, a representative sample was created with a deviation of the proportions of ethical groups from the target values within 0.00-0.04%. The statistical metrics obtained confirmed the effectiveness of the selected models and demonstrated a high degree of compliance with the ethical responsibility requirements of the results. The analysis showed that the initial datasets contained imbalances, which were successfully eliminated through multi-criteria optimization and data augmentation. The developed approach can be integrated into preparing training samples for ethically oriented artificial intelligence systems that perform automated detection of cyberbullying manifestations in text content, reducing the risks of reproducing social biases and increasing trust in algorithmic decisions.

A META-MODEL FOR LOW-CODE CONFIGURATION AND DEPLOYMENT OF CONTENT-BASED IMAGE RETRIEVAL SYSTEMS

Stanislav DANYLENKO , Serhii SMELYAKOV — Thu, 25 Sep 2025 00:00:00 +0300

The object of this study is content-based image retrieval (CBIR) systems with configurable architectures, and the subject is the meta-model for low-code configuration and deployment of CBIR systems. The goal of this work is to develop a meta-model for CBIR systems that enables their low-code configuration and deployment. The proposed meta-model allows users to define a CBIR system by selecting and combining the CBIR components from a predefined catalog, after which deployment artifacts are automatically generated and can be easily deployed. The proposed meta-model formalizes CBIR components: image repository, feature extractor, feature database (logical and physical levels), similarity measure, result aggregator, and user interaction layer; and extends them with two meta-level components: a configuration manager and a deployment engine. The architecture was implemented using Docker for containerization, Spring Boot starters for modularity, and a web-based graphical interface for configuration. A prototype was developed and tested on a dataset of 100 000 images, with systematic variation of component combinations. Experiments confirmed that the meta-model enables rapid reconfiguration and deployment of CBIR systems, allowing the evaluation of performance under different configurations. The examined difference between the best and worst tested configurations highlighting the significant effect of component selection on system performance. Scientific novelty lies in introducing a formalized meta-model that integrates low-code principles into CBIR design, combining modular architecture, containerized deployment, and graphical configuration in a single framework. The practical significance of the solution is in simplifying CBIR experimentation for researchers and practitioners without deep programming expertise, enabling rapid prototyping, testing, and deployment of customized CBIR systems.

MARKETING DECISION SUPPORT SYSTEM BASED ON FUZZY TRAINED ASSOCIATIVE RULES EXPERT SYSTEM

Thu, 25 Sep 2025 00:00:00 +0300

The fuzzy-associative metaheuristic approach addresses the urgent task of developing a marketing decision support system based on a fuzzy trained associative rules expert system, aimed at improving the accuracy and efficiency of consumer preference analysis. The proposed system combines the interpretability of fuzzy logic with data-driven learning via associative rules and parameter identification using an adaptive multi-agent optimization method. To achieve this goal, associative rule learning techniques (Apriori and FP-Growth) were used to extract frequent consumer behavior patterns. A fuzzy expert system was developed, in which the parameters of membership functions are optimized by the Adaptive Vibrating Particle System (AVPS) metaheuristic. Unlike traditional vibrating particle systems, AVPS integrates iteration-dependent control of particle positions, enabling global search in early iterations and local refinement at later stages, thus improving convergence speed and solution precision. The architecture was implemented using Python-based tools (TensorFlow, Keras, Pandas, mlxtend, Scikit-Fuzzy), and validated on the “Consumer Behavior and Shopping Habits” dataset. The fuzzy expert system achieved an accuracy of 0.98, outperforming human experts (0.80), traditional VPS optimization (0.93), and backpropagation-based training (0.90). The system also reduces reliance on manually tuned parameters and increases robustness to data incompleteness and noise. Scientific novelty lies in combining a fuzzy associative rule-learning framework with AVPS-based optimization, offering a scalable and interpretable decision-making mechanism. The developed system contributes to the advancement of intelligent recommendation engines, personalized marketing tools, and decision support systems in consumer-oriented analytics.

ANALYSIS OF EFFICIENCY OF HARDWARE PLATFROMS FOR SPATIAL ORIENTATION SYSTEMS USING A UNIFIED ENERGY CONSUMPTION MODEL

Oleksandr MAMCHYCH, Olesia BARKOVSKA, Andriy KOVALENKO, Anton HAVRASHEKNO — Thu, 25 Sep 2025 00:00:00 +0300

This paper addresses the problem of evaluating the efficiency of hardware platforms for spatial orientation systems, with a specific focus on mobile and assistive technologies for visually impaired users. The primary purpose is to conduct a systematic comparison between two classes of compact computing devices - single-board computers and smartphones - using a unified model and method for forecasting energy consumption in distributed computing systems that was developed and validated in our previous research. The methodology integrates both simulation and experimental measurements to provide a reliable assessment of computational performance, energy efficiency, and subsystem contributions under conditions representative of real-world computer vision workloads. The chosen experimental task was based on object detection using the SSD MobileNetV1 neural network, applied to video stream processing with standardized preprocessing and postprocessing stages, enabling reproducible and cross-platform evaluation. Energy consumption was decomposed into idle, computing, and camera subsystems, with measurements obtained through controlled power supply instrumentation over extended periods to eliminate short-term deviations. Results show that Apple smartphones consistently outperform single-board computers in both computational power and energy efficiency, with CPUs delivering significantly higher throughput and lower overall energy consumption during real-time inference, while GPU acceleration via CoreML further amplifies this advantage. Smartphones also demonstrate superior thermal stability and lower idle consumption, though their advanced camera subsystems introduce additional energy costs not observed in simpler USB cameras used with single-board platforms. The experiments shown that running similar task smartphones were underloaded and had a room for running better models, unreachable for single-board computers. The overall conclusion emphasizes that for computer vision tasks in spatial orientation systems, even older-generation smartphones represent a more efficient and practical hardware base than the most advanced single-board computers, offering not only higher performance per unit of energy but also a richer set of integrated sensors and connectivity options. These findings underline the strategic importance of smartphones as the optimal hardware foundation for next-generation assistive technologies, while pointing to future research directions involving Android platforms and peripheral expansions for single-board devices