The role of functional activation in neural networks in the context of financial time series analysis

2025;
: pp. 299–309
Received: April 18, 2024
Revised: March 01, 2025
Accepted: March 06, 2025

Senyk A. P., Manziy O. S., Pelekh V. R., Futryk Y. V., Senyk Y. A.  The role of functional activation in neural networks in the context of financial time series analysis.  Mathematical Modeling and Computing. Vol. 12, No. 1, pp. 299–309 (2025)  

1
Lviv Polytechnic National University
2
Lviv Polytechnic National University
3
Lviv Polytechnic National University
4
Lviv Polytechnic National University
5
National Forestry University of Ukraine

Nowadays, neural networks are among the most popular analysis tools.  They are effective in solving classification, pattern recognition, and clustering problems.  This paper provides a detailed description and analysis of the operational principles of two neural networks, namely a Siamese network and a multilayer perceptron.  A model for using these neural networks in time series forecasting is proposed.  As an example, a web application was created in which the described neural networks were used to analyze the correlation between pairs of financial assets and assess the risk level of an investment portfolio.  Modern information technologies, visualization methods, and advanced analysis tools used in the developed software product provide users with a comprehensive understanding of their investments, allowing them to assess risks and opportunities, as well as determine strategies for maximizing income and diversifying their selected set of financial assets.  The research results demonstrate the effectiveness of the Siamese network and multilayer perceptron in forecasting the prices of financial assets on the stock market and applying the obtained results in investment management tasks.

  1. Chicco D.  Siamese Neural Networks: An Overview.  Artificial Neural Networks. Methods in Molecular Biology.  2190, 73–94 (2021).
  2. Serrano N., Bellogín A.  Siamese neural networks in recommendation.  Neural Computing and Applications.  35, 13941–13953 (2023).
  3. Basu T., Menzer O., Ward J., SenGupta I.  A Novel Implementation of Siamese Type Neural Networks in Predicting Rare Fluctuations in Financial Time Series.  Risks.  10, 39 (2022).
  4. Malykhina G., Guseva A.  Application of Siamese Neural Networks for the Type of Emergency Determination.  CSIS'2019: Proceedings of the XI International Scientific Conference Communicative Strategies of the Information Society. 9, 1–5 (2020).
  5. Sverstiuk A., Dubynyak T., Manziy O., Senyk A., Ohloblin P.  Bayesian click model and methods of estimating its parameters.  CEUR Workshop Proceedings.  3628, 389–403 (2024).
  6. Dubyniak T. S., Manziy O., Gancarczyk T., Senyk A., Futryk Y.  Specialized information system for support of the process of recruiting securities.  CEUR Workshop Proceedings.  3468, 117–125 (2023).
  7. Senyk A., Manziy O., Futryk Y., Stepanyuk O., Senyk Y.  Information system supporting decision–making processes for forming of securities portfolio.  Journal of Lviv polytechnic national university "Information systems and networks".   11, 39–55 (2022).
  8. Senyk A. P., Manziy O. S., Ohloblin P. E., Senyk Y. A., Krasiuk O. P.  Application of the Bayesian approach to modeling credit risks.  Mathematical Modeling and Computing.  11 (4), 1025–1034 (2024).
  9. Stepanyuk O., Senyk A., Manziy O., Pavlyuk N., Senyk Y.  Information system of analysis for segmentation of the securities portfolio of the product market.  Scientific Messenger of LNU of Veterinary Medicine and Biote1chnologies. Series Economical Sciences.  25 (101), 11–21 (2023).
  10. Xu Z., Zhang J., Wang J., Xu Z.  RETRACTED ARTICLE: Prediction research of financial time series based on deep learning.  Soft Computing.  24, 8295–8312 (2020).
  11. Rezaee Z., Aliabadi S., Dorestani A., Rezaee N. J.  Application of Time Series Models in Business Research: Correlation, Association, Causation.  Sustainability.  12 (12), 4833 (2020).
  12. Brandt P. T., Williams J. T.  Multipletime series models. SAGE Publications, Inc. (2007).
  13. Goel A., Goel A. K., Kumar A.  The role of artificial neural network and machine learning in utilizing spatial information.  Spatial Information Research.  31 (3), 275–285 (2023).
  14. Parhi R., Nowak R. D.  The Role of Neural Network Activation Functions.  IEEE Signal Processing Letters.  27, 1779–1783 (2020).