Перегляд за Автор "Levashenko, Vitaly"

Зараз показуємо 1 - 2 з 2
  • Ескіз
    Документ
    Machine Learning Models for Predicting Migrant Remittance Flows: A Cross-Border Financial Analysis
    (CEUR, 2026-02-07) Dumanska, Ilona; Kuzmin, Andrii; Levashenko, Vitaly; Lysak, Viktor; Hrytsyna, Lesia
    The paper proposes a machine learning framework for forecasting migrant remittance flows, focusing on the Ukraine–Poland corridor during 2022–2025. The methodology integrates diverse data sources— migration volumes, conflict intensity indices, exchange rates, host-country employment rates, and social sentiment—into a unified time-series dataset. Four models are evaluated: Linear Regression (baseline), Random Forest, XGBoost, and LSTM. LSTM is expected to outperform others due to its ability to capture long-term dependencies and crisis-driven shocks. Feature-importance analysis will likely highlight migration volume, employment rate, and exchange rate as key predictors, while sentiment data should enhance short-term responsiveness. The case study illustrates how remittance flows correlate with refugee inflows, labor integration, and policy interventions. Overall, the framework shows the potential of deep learning and ensemble methods to improve forecasting under humanitarian and economic stress.
  • Ескіз
    Документ
    Method and cyber-physical system for forecasting and optimizing electricity consumption in residential districts
    (Хмельницький національний університет, 2025) Pysmeniuk, Volodymyr; Levashenko, Vitaly
    The development of cyber-physical systems combined with machine learning algorithms opens new opportunities for forecasting and optimizing electricity consumption in residential districts. This study examined existing technologies and solutions for energy consumption management, identifying their advantages and disadvantages. The analysis showed that modern commercial systems are primarily designed either for industrial use or individual consumption, lacking a comprehensive approach for residential districts. The proposed forecasting and optimization method is based on hybrid machine learning algorithms. For energy consumption forecasting, a combination of recurrent neural networks (RNN) and XGBoost was used, allowing for the consideration of both temporal dependencies and nonlinear factors. For energy consumption optimization, a combination of genetic algorithms (GA) and particle swarm optimization (PSO) was implemented, ensuring efficiency in finding optimal solutions. The developed cyber-physical system includes sensors for data collection, microcontrollers (Raspberry Pi) for data processing, and intelligent systems for controlling electrical appliances. This enables real-time energy consumption analysis and management, improving the energy efficiency of residential districts. Experimental results confirmed the effectiveness of the proposed approach, demonstrating high accuracy in energy consumption forecasting and the potential for reducing electricity costs through optimized usage. The proposed method has significant potential for scaling and implementation in large residential complexes, contributing to sustainable development and reducing the load on energy grids. Thus, the results of this study can be used for further improvement of energy management systems, promoting efficient electricity use, reducing consumer costs, and minimizing the environmental impact of energy systems.