Browsing by Author "Ostrovska, Kateryna Yu."

Now showing 1 - 6 of 6

Application of Neural Networks for Prediction Financial Time Series
(Scientific Publishing Center “Sci-conf.com.ua”, Perfect Publishing, 2024) Prokofiev, Taras; Ostrovska, Kateryna Yu.
ENG: The article discusses some aspects and features of the use of neural networks for forecasting financial time series for the purpose of making a profit. The use of neural networks to analyze financial information is a promising alternative (or complement) to traditional research methods. Due to their adaptability, the same neural networks can be used to analyze several instruments and markets, while the patterns found by a player for a specific instrument using technical analysis methods may work worse or not work at all for other instruments.
Development and Research of a Chatbot Using the Linguistic Core of Amazon Lex V2
(CEUR-WS Team, Aachen, Germany, 2024) Hnatushenko, Viktoriia V.; Ostrovska, Kateryna Yu.; Nosov, Valerii
ENG: The main of this research is to develop and explore the configuration of a text and voice recognition system, integrate it into a specialized application, and deploy the application in a cloud environment. Amazon Lex service is built on chatbots that support Natural Language Understanding (NLU) and voice recognition. The developed chatbot elevates the user experience while engaging with voice consultants by offering flexible customization options. A chatbot has been designed with interactive text input fields and voice recording functions. The server architecture of the application is configured for seamless data transmission through the AWS SDK to Amazon Lex. The input information undergoes processing to ensure the generation of responses that are dynamically displayed on the web page. The structure of all intents – simulating banking services such as checking card balance, transaction history, and more. Testing the intents was done by creating a dataset with possible user statements and automated runs. The developed chatbot was tested through 6 runs, each consisting of up to 5 statements for recognition. The accuracy of text input recognition ranged from 60% to 99%, with voice input recognition accuracy being 10% lower.
Machine Learning Methods for Antifraud Systems
(Український державний університет науки і технологій, ННІ ≪Дніпровський металургійний інститут≫, ІВК ≪Системні технології≫, Дніпро, 2025) Ostrovska, Kateryna Yu.; Nosov, Valerii O.
ENG: Fraud in the financial sector, e-commerce, and online services is becoming increasingly frequent and sophisticated. Traditional rule-based systems, while still helpful in detecting known fraud patterns, struggle to keep up with new, evolving attack vectors, as static rules are quickly circumvented. In contrast, machine learning (ML) provides a dynamic and scalable approach that can process vast amounts of transactional and behavioral data to identify subtle anomalies and suspicious activity. This paper provides a comprehensive overview of current ML techniques used in fraud detection, categorized into three main groups: classification models, anomaly detection methods, and deep learning architectures. It discusses real-world applications across various fraud scenarios, including credit card abuse, account takeovers, cybercrime, and scams in digital comerce. Emphasis is placed on the strengths and limitations of each approach, with attention to real-world considerations like scalability, model transparency, and the challenge of class imbalance. The paper also reviews recent advances, including graph-based representations of financial interactions, IP-based behavioral profiling, and the emergence of hybrid systems that integrate multiple ML techniques –such as combining autoencoders with boosting algorithms for improved accuracy, especially when labeled data is scarce. The findings aim to support the development of flexible, high-performance fraud detection solutions that leverage the most effective ML practices and capitalize on the synergy of hybrid model architectures.
Research of Image Classification Methods Using Neural Networks on GPUs
(Scientific Publishing Center “Sci-conf.com.ua”, Perfect Publishing, 2024) Cherskyi, Serhii; Ostrovska, Kateryna Yu.
ENG: The paper examines the classification of images on GPUs by means of neural networks, namely, using the example of the categorization of household goods. This topic is relevant, since in everyday life we are surrounded by images and it is easy for a person to interpret them, and it is much more difficult for a computer, all the more to classify or segment images. As a result, a system was created that automatically classifies goods, modifying existing approaches, and obtained a custom one that works better for this task. Having improved the product, it can be used for any organization where it would be convenient to automatically classify products.
Research of Image Classification Methods Using Neural Networks on GPUs
(Scientific Publishing Center “Sci-conf.com.ua”, BoScience Publisher, 2024) Cherskyi, Serhii; Ostrovska, Kateryna Yu.
ENG: The paper examines the classification of images on GPUs by means of neural networks, namely, using the example of the categorization of household goods. This topic is relevant, since in everyday life we are surrounded by images and it is easy for a person to interpret them, and it is much more difficult for a computer, all the more to classify or segment images. As a result, a system was created that automatically classifies goods, modifying existing approaches, and obtained a custom one that works better for this task. Having improved the product, it can be used for any organization where it would be convenient to automatically classify products.
Research on Environmental Changes Based on Fractal Characteristics of Satellite Images
(CEUR-WS Team, Aachen, Germany, 2025) Hnatushenko, Viktoriia V.; Zhurba, Anna O.; Zimoglyad, Andrew Yu.; Ostrovska, Kateryna Yu.
ENG: Most natural structures that are widely studied today using computer science have a complex fractal structure. Fractal analysis of such structures is used to model, study, and explain the properties of surfaces and structures of complex objects in various fields of science and technology. Images of a large number of natural surfaces and structures are satellite images that exhibit fractal properties. Satellite images in modern life have high spatial resolution, which gives researchers and users satisfactory initial data for solving various types of problems. A promising direction for increasing the informativeness of satellite images is the use of fractal image analysis methods. The complexity of the forms of the underlying surface and vegetation can be described using the fractal dimension. Characteristic values of the fractal dimension allow decoding of space images. The paper proposes a method for studying environmental changes in satellite images based on the calculation of fractal characteristics, such as fractal dimension, fractal distribution and fractal segmentation. The ecological indicators for assessing the state of the environment were selected as trends in forest numbers, water and land resources. Satellite images of the Amazon forests, Bolivia in 1991, 1996, 2006, 2012, 2016, 2020, which were subjected to mass deforestation, were selected for the study. The experimental results show that the fractal dimension increases each time (Fractal Dimension (FR) = 1.441 in 1991, FR = 1.825 in 2020), and the green areas in this area decrease. The depth of the seas (Black Sea, Tyrrhenian Sea, Mediterranean Sea, Philippine Sea) was studied. The least homogeneous with the largest amplitude of distribution modes has the fractal distribution of the Philippine Sea, which indicates a more pronounced relief of the seabed. As a result of the study of winter fields with different levels of snow cover, it was found that an increase in its value leads to an increase in the value of the fractal dimension (FR=1.702 February, FR=1.894 March). Thus, fractal analysis of winter fields allows us to estimate the relative amount of moisture that will enter the soil in the spring. The study highlights the need for further research in developing more efficient fractal methods to improve the accuracy of change area detection, which will favor the analysis of the causes and consequences of the environmental situation.