Igor B. Olenych
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Igor B. Olenych has a total of 83 co-authors. Here are the top co-authors by number of shared publications:
L. S. Monastyrskii
12 shared publications- Andriy Luchechko
Ivan Franko National University of Lviv
8 shared publications O. I. Aksimentyeva
8 shared publications- Olena Aksimentyeva
Department of Physical and Colloidal Chemistry, Ivan Franko National University of Lviv
7 shared publications - Bogdan S. Sokolovskii
Faculty of Electronics and Computer Technologies , Ivan Franko National University of Lviv , Lviv , Ukraine
6 shared publications Yu. Yu. Horbenko
Chemistry Department, Ivan Franko National University of Lviv, 79005 Lviv, Ukraine
6 shared publications
Latest Publications
- Open Access
- 31 October 2026
- Published by Private Company Technology Center in Technology audit and production reserves
- Vol. 5 (2(91)) , 29-37
Development of a framework for data volume impact assessment on the advantage of boosting methods over classical in streaming financial data forecasting
The object of research is the process of analyzing and predicting the characteristics of streaming data using machine learning methods. The problem to be solved is the lack of a methodologically rigorous approach to comparing machine learning models on streaming data. Existing researches often evaluate models using global metrics. That metrics do not take into account the heterogeneity of individual trading pairs and also overestimate the apparent quality of models. The results of comparisons are rarely checked for statistical significance, and the impact of data volume on the relative advantage of models remains poorly studied. A methodological framework is proposed for evaluating data volume effect on boosting methods advantage over classical methods in forecasting streaming financial data. Every trading pair in framework included computed metrics and statistical significance using Wilcoxon signed-rank test. 6 machine learning models were used for volume forecasting. 3 different scales were implemented for simultaneous stream data forecast from Apache Kafka. Their goal was to process 3.9 million, 19.6 million and 70 million records for 46, 91 and 148 trading pairs. Global metrics can substantially skew real situation. Naive approaches R2 = 0.9334 globally but only R2 = 0.2463 pairwise. Best performing models are LightGBM and XGBoost with R2 = 0.9543 and 0.9549 globally but only R2 = 0.4837 and 0.4899 pairwise. Wilcoxon signed-rank test approved statistical significance with p < 0.001. XGBoost outmatched Linear Regression in 97% of 148 pairs. Boosting advantage corresponds to data volume and increases with it. Framework applies open source software and can be recreated. Research results can be helpful in metrics and machine learning models choice for streaming financial data forecasting. The framework can be implemented in systems for forecasting cryptocurrency trading volume, volatility, and order execution optimization. It is a guideline for choosing tools in real-time financial systems.
- 31 October 2026
- Published by Private Company Technology Center in Technology audit and production reserves
- Vol. 5 (2(91)) , 29-37
Open AccessDevelopment of a framework for data volume impact assessment on the advantage of boosting methods over classical in streaming financial data forecasting
The object of research is the process of analyzing and predicting the characteristics of streaming data using machine learning methods. The problem to be solved is the lack of a methodologically rigorous approach to comparing machine learning models on streaming data. Existing researches often evaluate models using global metrics. That metrics do not take into account the heterogeneity of individual trading pairs and also overestimate the apparent quality of models. The results of comparisons are rarely checked for statistical significance, and the impact of data volume on the relative advantage of models remains poorly studied. A methodological framework is proposed for evaluating data volume effect on boosting methods advantage over classical methods in forecasting streaming financial data. Every trading pair in framework included computed metrics and statistical significance using Wilcoxon signed-rank test. 6 machine learning models were used for volume forecasting. 3 different scales were implemented for simultaneous stream data forecast from Apache Kafka. Their goal was to process 3.9 million, 19.6 million and 70 million records for 46, 91 and 148 trading pairs. Global metrics can substantially skew real situation. Naive approaches R2 = 0.9334 globally but only R2 = 0.2463 pairwise. Best performing models are LightGBM and XGBoost with R2 = 0.9543 and 0.9549 globally but only R2 = 0.4837 and 0.4899 pairwise. Wilcoxon signed-rank test approved statistical significance with p < 0.001. XGBoost outmatched Linear Regression in 97% of 148 pairs. Boosting advantage corresponds to data volume and increases with it. Framework applies open source software and can be recreated. Research results can be helpful in metrics and machine learning models choice for streaming financial data forecasting. The framework can be implemented in systems for forecasting cryptocurrency trading volume, volatility, and order execution optimization. It is a guideline for choosing tools in real-time financial systems.
- 4 September 2026
- Published by Taylor & Francis in Molecular Crystals and Liquid Crystals
- Vol. 770 (14) , 2497-2507
Field-effect photodetector with extended spectral sensitivity based on reduced graphene oxide
1. Carbon nanostructures have become key candidates as base material for next-generation electronics, stimulating innovation in many technological fields. Graphene attracts special attention from r...
- 4 September 2026
- Published by Taylor & Francis in Molecular Crystals and Liquid Crystals
- Vol. 770 (14) , 2497-2507
Field-effect photodetector with extended spectral sensitivity based on reduced graphene oxide
1. Carbon nanostructures have become key candidates as base material for next-generation electronics, stimulating innovation in many technological fields. Graphene attracts special attention from r...
- Open Access
- 20 April 2026
- Published by Odesa I.I. Mechnikov National University in Sensor Electronics and Microsystem Technologies
- Vol. 23 (1) , 27-34
OPTIMIZATION OF THE STRUCTURE OF FIELD EFFECT TRANSISTORS BASED ON REDUCED GRAPHENE OXIDE FILM FOR PHOTOAND GAS-SENSING APPLICATIONS
The features of using field-effect transistors (FETs) based on a reduced graphene oxide (RGO) film deposited on the surface of anodically and thermally oxidized porous silicon as photodetectors and gas adsorption sensors have been investigated. An increase in the efficiency of the created FETs was found as a result of thermal oxidation of the porous layer or deposition of an additional SiO₂ layer on the surface of anodically oxidized porous silicon. It has been established that the photo- and adsorption sensitivity of the RGO-based FETs depends on both the quality of the insulating layer and the presence of a protective SiO₂ coating on the RGO film. In particular, the FETs with an anodic oxidized porous silicon insulating layer are characterized by the highest photosensitivity. Conversely, the graphene FET with thermally oxidized porous silicon is almost insensitive to visible light but demonstrates maximum response to the adsorption of water molecules. The SiO₂ protective layer has almost no effect on the photosensitivity of RGO-based FETs but significantly reduces the influence of the atmosphere on their conductivity.
- 20 April 2026
- Published by Odesa I.I. Mechnikov National University in Sensor Electronics and Microsystem Technologies
- Vol. 23 (1) , 27-34
Open AccessOPTIMIZATION OF THE STRUCTURE OF FIELD EFFECT TRANSISTORS BASED ON REDUCED GRAPHENE OXIDE FILM FOR PHOTOAND GAS-SENSING APPLICATIONS
The features of using field-effect transistors (FETs) based on a reduced graphene oxide (RGO) film deposited on the surface of anodically and thermally oxidized porous silicon as photodetectors and gas adsorption sensors have been investigated. An increase in the efficiency of the created FETs was found as a result of thermal oxidation of the porous layer or deposition of an additional SiO₂ layer on the surface of anodically oxidized porous silicon. It has been established that the photo- and adsorption sensitivity of the RGO-based FETs depends on both the quality of the insulating layer and the presence of a protective SiO₂ coating on the RGO film. In particular, the FETs with an anodic oxidized porous silicon insulating layer are characterized by the highest photosensitivity. Conversely, the graphene FET with thermally oxidized porous silicon is almost insensitive to visible light but demonstrates maximum response to the adsorption of water molecules. The SiO₂ protective layer has almost no effect on the photosensitivity of RGO-based FETs but significantly reduces the influence of the atmosphere on their conductivity.
- Open Access
- 28 February 2026
- Published by Institute of Physical Optics in Ukrainian Journal of Physical Optics
- Vol. 27 (3) , 03001-03009
COLOR-TUNABLE PHOTOEMISSION IN AlQ3-POROUS SILICON HYBRID STRUCTURES
- 28 February 2026
- Published by Institute of Physical Optics in Ukrainian Journal of Physical Optics
- Vol. 27 (3) , 03001-03009
Open AccessCOLOR-TUNABLE PHOTOEMISSION IN AlQ3-POROUS SILICON HYBRID STRUCTURES
- Open Access
- 29 January 2026
- Published by Khmelnytskyi National University in Herald of Khmelnytskyi National University. Technical sciences
- Vol. 361 (1) , 383-389
COMPARATIVE STUDY OF MACHINE LEARNING METHODS FOR STREAMING DATA PROCESSING
The process of analysing large-scale streaming cryptocurrency data by machine learning algorithms is the object of this research. Handling terabyte-scale, high-velocity data streams presents a critical challenge due to the computational and accuracy limitations of classical machine learning methods, which struggle with the volume and complexity of millions of temporal records. The principal result is the development of a distributed processing pipeline featuring a Feature Store architecture. This solution enabled LightGBM and XGBoost algorithms to achieve superior predictive performance (R² was 0.9998 and 0.9997, respectively) while processing 1.33 million streaming records across 100 cryptocurrency pairs. The research methodology included a comprehensive feature engineering phase, extracting a set of temporal, statistical, and technical indicators, such as rolling means, volatility measures, and lagged price values, which are crucial for capturing dependencies in big data. This performance advantage is attributed to the architectural capabilities of gradient boosting algorithms. The proposed pipeline successfully shifts the process from conventional linear approaches to advanced tree-based ensemble methods with optimized memory management, demonstrating that gradient boosting algorithms possess the necessary computational efficiency and pattern recognition capabilities that Decision Tree, Random Forest, and Regression methods lack. In practice, the findings provide clear guidelines for big data practitioners. The Feature Store architecture with temporal stratified sampling is a scalable framework achieving 5.7x data reduction and near 82% memory savings. For production systems handling high-velocity streaming data, gradient boosting algorithms (particularly LightGBM with 0.63 s training time) are the superior strategy over traditional methods for achieving both accuracy and computational efficiency.
- 29 January 2026
- Published by Khmelnytskyi National University in Herald of Khmelnytskyi National University. Technical sciences
- Vol. 361 (1) , 383-389
Open AccessCOMPARATIVE STUDY OF MACHINE LEARNING METHODS FOR STREAMING DATA PROCESSING
The process of analysing large-scale streaming cryptocurrency data by machine learning algorithms is the object of this research. Handling terabyte-scale, high-velocity data streams presents a critical challenge due to the computational and accuracy limitations of classical machine learning methods, which struggle with the volume and complexity of millions of temporal records. The principal result is the development of a distributed processing pipeline featuring a Feature Store architecture. This solution enabled LightGBM and XGBoost algorithms to achieve superior predictive performance (R² was 0.9998 and 0.9997, respectively) while processing 1.33 million streaming records across 100 cryptocurrency pairs. The research methodology included a comprehensive feature engineering phase, extracting a set of temporal, statistical, and technical indicators, such as rolling means, volatility measures, and lagged price values, which are crucial for capturing dependencies in big data. This performance advantage is attributed to the architectural capabilities of gradient boosting algorithms. The proposed pipeline successfully shifts the process from conventional linear approaches to advanced tree-based ensemble methods with optimized memory management, demonstrating that gradient boosting algorithms possess the necessary computational efficiency and pattern recognition capabilities that Decision Tree, Random Forest, and Regression methods lack. In practice, the findings provide clear guidelines for big data practitioners. The Feature Store architecture with temporal stratified sampling is a scalable framework achieving 5.7x data reduction and near 82% memory savings. For production systems handling high-velocity streaming data, gradient boosting algorithms (particularly LightGBM with 0.63 s training time) are the superior strategy over traditional methods for achieving both accuracy and computational efficiency.
- Open Access
- 1 January 2026
- Published by Sumy State University in Journal of Nano- and Electronic Physics
- Vol. 18 (2) , 02026-1
Field-Effect Transistor Based on Reduced Graphene Oxide Film with ZnO and Porous Silicon Absorbing Layers for Ionizing Radiation Detection
Journal of Nano- and Electronic Physics. Scientific journal. ISSN: 2077-6772. Journal abbreviation: J. Nano- Electron. Phys. No page charges. All articles are freely available on-line. Issued 4 times per year. Publisher: Sumy State University. (Sumy, Ukraine)
- 1 January 2026
- Published by Sumy State University in Journal of Nano- and Electronic Physics
- Vol. 18 (2) , 02026-1
Open AccessField-Effect Transistor Based on Reduced Graphene Oxide Film with ZnO and Porous Silicon Absorbing Layers for Ionizing Radiation Detection
Journal of Nano- and Electronic Physics. Scientific journal. ISSN: 2077-6772. Journal abbreviation: J. Nano- Electron. Phys. No page charges. All articles are freely available on-line. Issued 4 times per year. Publisher: Sumy State University. (Sumy, Ukraine)
- Open Access
- 1 January 2026
- Published by Walter de Gruyter GmbH in Applied Computer Systems
- Vol. 31 (1) , 63-73
Physics-Inspired Hamiltonian Particle Swarm Optimisation for Multi-Agent Movement
This study puts forward a Hamiltonian-inspired modification of Particle Swarm Optimisation (PSO) algorithm. Since the standard PSO procedure does not take into account physical properties like particle masses, geometrical sizes, and energy consumption, it is not fully applicable as a navigational and coordination tool in real-world environments. In particular, generic PSO mechanism cannot stop the particles from collisions. To address these issues, we propose a new PSO formulation based on a Hamiltonian interpretation. This approach allows bringing together the kinetic and potential energy terms with the forces acting on the agents, as well as the derivation of agent’s velocities and positions. The potential energy represents attraction toward both personal and global best positions in a spring-like manner. As a component of conservative forces derived from the potential energy term, we introduce a special repulsive potential function to prevent collisions among agents. The kinetic energy, which is derived via agent mass and momentum, determines the movement dynamics. To model the energy loss, we incorporate Rayleigh dissipation term that accounts for non-conservative forces. According to the proposed model, agent displacements are computed using the obtained velocity and momentum vectors. Additionally, we introduce individual and swarm energy efficiency metrics to study the agents’ motion in a 2D testing environment. The presented approach enables stable, coordinated, and collision-free multi-agent motion within a physics-inspired optimisation framework.
- 1 January 2026
- Published by Walter de Gruyter GmbH in Applied Computer Systems
- Vol. 31 (1) , 63-73
Open AccessPhysics-Inspired Hamiltonian Particle Swarm Optimisation for Multi-Agent Movement
This study puts forward a Hamiltonian-inspired modification of Particle Swarm Optimisation (PSO) algorithm. Since the standard PSO procedure does not take into account physical properties like particle masses, geometrical sizes, and energy consumption, it is not fully applicable as a navigational and coordination tool in real-world environments. In particular, generic PSO mechanism cannot stop the particles from collisions. To address these issues, we propose a new PSO formulation based on a Hamiltonian interpretation. This approach allows bringing together the kinetic and potential energy terms with the forces acting on the agents, as well as the derivation of agent’s velocities and positions. The potential energy represents attraction toward both personal and global best positions in a spring-like manner. As a component of conservative forces derived from the potential energy term, we introduce a special repulsive potential function to prevent collisions among agents. The kinetic energy, which is derived via agent mass and momentum, determines the movement dynamics. To model the energy loss, we incorporate Rayleigh dissipation term that accounts for non-conservative forces. According to the proposed model, agent displacements are computed using the obtained velocity and momentum vectors. Additionally, we introduce individual and swarm energy efficiency metrics to study the agents’ motion in a 2D testing environment. The presented approach enables stable, coordinated, and collision-free multi-agent motion within a physics-inspired optimisation framework.
- 15 December 2025
- Published by Ivan Franko National University of Lviv in Journal of Physical Studies
- Vol. 29 (4)
Charge carrier mobility in FETs based on reduced graphene oxide film with RF magnetron sputtered dielectric layers
- 15 December 2025
- Published by Ivan Franko National University of Lviv in Journal of Physical Studies
- Vol. 29 (4)
Charge carrier mobility in FETs based on reduced graphene oxide film with RF magnetron sputtered dielectric layers
- 30 July 2025
- Published by Taylor & Francis in Molecular Crystals and Liquid Crystals
- Vol. 769 (15-18) , 1527-1537
Field-effect transistor based on reduced graphene oxide – porous silicon hybrid structure for application in e-nose technology
- 30 July 2025
- Published by Taylor & Francis in Molecular Crystals and Liquid Crystals
- Vol. 769 (15-18) , 1527-1537
Field-effect transistor based on reduced graphene oxide – porous silicon hybrid structure for application in e-nose technology
- Open Access
- 23 June 2025
- Published by Odesa I.I. Mechnikov National University in Sensor Electronics and Microsystem Technologies
- Vol. 22 (2) , 19-26
FIELD-EFFECT TRANSISTORS BASED ON REDUCED GRAPHENE OXIDE FILM FOR PHOTO AND RADIATION DETECTORS
The features of using dielectric layers of SiO2, Al2O3, and HfO2 in field-effect transistors based on reduced graphene oxide (RGO) films for detecting visible and ionizing radiation have been investigated. The influence of the dielectric layer on the electrical characteristics of the created field-effect transistors in DC and AC modes was studied based on an analysis of the dependencies of drain current and RGO film resistance on gate voltage. An increase in drain current, a decrease in RGO film resistance, and a shift of the charge neutrality point toward negative gate voltages were found under the influence of white light irradiation. It was established that irradiation with β-particles and γ-quanta from a 226Ra source causes a transformation of the capacitance-voltage characteristics of the obtained field-effect transistors based on the RGO film due to the formation of electrically active defects in the dielectric layer. The effectiveness of using various dielectric layers to create photo- and ionizing radiation detectors based on graphene field-effect transistors is discussed.
- 23 June 2025
- Published by Odesa I.I. Mechnikov National University in Sensor Electronics and Microsystem Technologies
- Vol. 22 (2) , 19-26
Open AccessFIELD-EFFECT TRANSISTORS BASED ON REDUCED GRAPHENE OXIDE FILM FOR PHOTO AND RADIATION DETECTORS
The features of using dielectric layers of SiO2, Al2O3, and HfO2 in field-effect transistors based on reduced graphene oxide (RGO) films for detecting visible and ionizing radiation have been investigated. The influence of the dielectric layer on the electrical characteristics of the created field-effect transistors in DC and AC modes was studied based on an analysis of the dependencies of drain current and RGO film resistance on gate voltage. An increase in drain current, a decrease in RGO film resistance, and a shift of the charge neutrality point toward negative gate voltages were found under the influence of white light irradiation. It was established that irradiation with β-particles and γ-quanta from a 226Ra source causes a transformation of the capacitance-voltage characteristics of the obtained field-effect transistors based on the RGO film due to the formation of electrically active defects in the dielectric layer. The effectiveness of using various dielectric layers to create photo- and ionizing radiation detectors based on graphene field-effect transistors is discussed.