INFORMATION SYSTEM FOR FORECASTING NONLINEAR NON-STATIONARY PROCESSES IN FINANCE
DOI:
https://doi.org/10.20535/kpisn.2024.1-4.312719Abstract
Background. Financial processes are often characterised by nonlinearity and non-stationarity, which makes them difficult to accurately model and forecast. Traditional methods cannot effectively take into account the complex interrelationships and variability of such processes, which generates increased uncertainty and risks. This leads to the need to develop new information systems and methods to improve the accuracy and sustainability of forecasts.
Objective. To provide a brief overview of the characteristics of nonlinear non-stationary processes, to develop a methodology for their modelling, as well as to build mathematical models based on actual statistical data and to obtain practically useful results of modelling and forecasting selected processes.
Methods. The methodology for modelling and forecasting nonlinear non-stationary processes is applied, models are built using data mining, such as regression models and a neural network, and the main metrics for assessing the adequacy of the model and quality of the forecast are used.
Results. The developed information system for modelling and forecasting nonlinear non-stationary processes is approbated on real statistical data. Based on data mining methods, models of the share price dynamics of a well-known company were built. The study's results demonstrate that using an integrated approach, which includes regression models and neural networks, significantly improves the quality of forecasting variance changing in time and the nonlinear non-stationary process.
Conclusions. The task of high-quality forecasting of processes due to rapid, sometimes hard-to-predict changes in the external environment, i.e. external shocks, which is typical for nonlinear non-stationary financial processes, is still relevant today. The literature provides a sufficient variety of methods for modelling these processes. However, in this research, the methods that have demonstrated their advantages in modelling financial transactions in the stock market were chosen, and therefore it makes sense to expand and improve the perspectives of this approach.
References
C. Cheng, et al. Time series forecasting for nonlinear and non-stationary processes: a review and comparative study. Iie Transactions, 2015, pp. 1053-1071.
A. Hasanzadeh, et al. Piecewise stationary modeling of random processes over graphs with an application to traffic prediction. In: 2019 IEEE International Conference on Big Data (Big Data). IEEE, 2019, pp. 3779-3788.
Y. Chang, J. Y. Park, P. C. Phillips. Nonlinear econometric models with cointegrated and deterministically trending regressors. The Econometrics Journal, 4.1, 2001, pp. 1-36.
J. Y. Park. Nonstationary nonlinearity: An outlook for new opportunities. mimeographed, Deparment of Economics, Rice University, pp. 21-23, 2003.
P.I. Bidyuk, L.O. Korshevnyuk. Design of computer information systems for decision support. Kyiv: Educational book, ESC "IASA" NTUU "KPI", 2010, p. 340.
K. Turkman, M. G. Scotto, Z. B. Patrícia. Non-linear time series (Vol. 30). Switzerland: Springer Publications, 2014, pp. 121-193.
H. Tsai, K. S. Chan. Testing for nonlinearity with partially observed time series. Great Britain: Biometrika, 2000, pp. 805-821.
C. He, R. Sandberg. Dickey–Fuller type of tests against nonlinear dynamic models. Stockholm: Oxford Bulletin of Economics and Statistics, 2006, pp. 835-861.
A. B. Abdul-Hameed, O. G. Matanmi. A modified Breusch–Pagan test for detecting heteroskedasticity in the presence of outliers. Nigeria: Pure and Applied Mathematics Journal, 2021, pp. 139-149.
J. A. Visek. Empirical distribution function under heteroscedasticity. Czech Republic: Statistics, 45(5), 2011, pp. 497-508.
S. Bianconcini, E. B. Dagum, P. Maass, T. Alexandrov, T. S. McElroy. A review of some modern approaches to the problem of trend extraction. Washington: Econometric Reviews, 31(6), 2012, pp. 593-624.
T. Alexandrov. A method of trend extraction using singular spectrum analysis. Germany: Center for Industrial Mathematics, 2008, pp. 1-23.
J. M. Dufour, M. G. Dagenais. Durbin-Watson tests for serial correlation in regressions with missing observations. Canada: Journal of Econometrics, 27(3), 1985, pp. 371-381.
J. E. Cavanaugh, A. A. Neath. The Akaike information criterion: Background, derivation, properties, application, interpretation, and refinements. Iowa: Wiley Interdisciplinary Reviews, Computational Statistics, 11(3), 2019, pp. 1-11.
T. O. Hodson. Root mean square error (RMSE) or mean absolute error (MAE): When to use them or not. USA: Geoscientific Model Development Discussions, 2022, pp. 5481-5487.
P.I. Bidyuk, et al. Adaptive modeling and forecasting economic and financial processes. Ukraine: Informatics and Mathematical Methods in Simulation
Vol. 9, No. 4, 2019, pp. 231-250.
J. Fan, Q. Yao. Nonlinear Time Series: Nonparametric and Parametric Methods. New York: Springer, 2003, pp. 143—179.
R.S. Tsay, Analysis of financial time series. Chicago: Wiley & Sons, Ltd., 2010, pp. 115—140.
A. Sagheer, M. Kotb, Time series forecasting of petroleum production using deep LSTM recurrent networks. Neurocomputing, 2019, pp. 203-213.
S. Siami-Namini, N. Tavakoli, A. S. Namin. A comparison of ARIMA and LSTM in forecasting time series. Texas: International conference on machine learning and applications (ICMLA), 2018. pp. 1394-1401.
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Petro Bidyuk, Roman Tertychnyi, Denys Chuprin
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under CC BY 4.0 that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work
