Vitaliy Kobets, Professor of the Department of Computer Science and Software Engineering, Participated in the VI International Scientific and Practical Conference
Vitaliy Kobets, Professor of the Department of Computer Science and Software Engineering, participated in the VI International Scientific and Practical Conference for Scientists and Students: "Digital Economy as a Factor of Innovation and Sustainable Development of Society." The event, held on December 9–10, 2025, at the Ivan Puluj Ternopil National Technical University, brought together researchers, educators, practitioners, and young scientists to present cutting-edge developments in digitalization, innovation, and sustainable development.
During the plenary session, Professor Kobets delivered a presentation titled "Financial Video Analysis Using Natural Language Processing: An Empirical Study of Stock Price Prediction." He presented research on the application of NLP (Natural Language Processing) in forecasting stock prices compared to traditional methods, such as ARIMA (AutoRegressive Integrated Moving Average).
Key findings from the study include:
-
Accuracy: When comparing forecasts with actual stock price fluctuations, the system achieved 85% overall accuracy. This confirms the viability of NLP-based automation for financial video analysis, surpassing the accuracy of traditional time-series models.
-
Strategic Value: The results highlight the potential of such systems to support investment decision-making.
-
Identified Limitations: The study noted challenges in processing long-term market volatility, imperfections in interpreting neutral forecasts, and a lack of multimodal processing for visual cues (such as charts or non-verbal signals).
The research concludes that the NLP approach outperforms ARIMA time-series models for the financial instruments analyzed.
Future Research Directions: Future studies will focus on developing multimodal models that integrate text, audio, and visual signals, alongside expanding validation to more diverse datasets. These advancements are expected to enhance contextual accuracy, reduce bias, and increase the reliability of automated financial forecasting based on unstructured media sources.