SHORT-TERM OF NON-STATIONARY TIME SERIES FORECASTING USING MLP AND LSTM MODELS
DOI:
https://doi.org/10.26906/SUNZ.2023.1.091Keywords:
time series forecasting, computational intelligence, artificial neural network, multilayer perceptron, long shortterm memory, M3-CompetitionAbstract
Topicality. The solution of the problem of forecasting takes an important role in the processes of strategic planning and operational management in various areas of economic activity. A widespread form of forecasting is time series forecasting, in which the actual problem is the choice of an appropriate method among modern means of computational intelligence, such as artificial neural networks. The problem of choice is due to a large number of parameters and settings that depend on the features of the predicted time series and have a significant impact on the forecast quality. The goal of this work is to analyze methods for short-term forecasting of non-stationary time series using artificial neural network models such as multilayer perceptron and long short-term memory. The object of research is the process of time series forecasting. The subject of research is the use of artificial neural network models for short-term forecasting. Results. Experimental research has shown that the average error in forecasting using the proposed tools is 2-6% lower compared to using common traditional models. Conclusions. The obtained results confirm the advantage of using the MLP and LSTM models over forecasting based on the methods chosen for the M3- Competition analysis.Downloads
References
Iorio C., Frasso G., Ambrosio A. (2016), “Parsimonious time series clustering using P-splines”, Expert Systems with Applications, № 52, pp. 26-38.
Yang J. (2006), “Power System Short-term Load Forecasting: Thesis for Ph.d degree”, Elektrotechnik und Informationstechnik der Technischen Universitat, Darmstadt, Germany, P. 139.
Korablev N.M., Ivaschenko G.S. (2014), “Parallel immune algorithm of short-term forecasting based on model of clonal selection”, Radio Electronics, Computer Science, Control, № 2. pp. 73-78, doi: doi.org/10.15588/1607-3274-2014-2-11
Ivaschenko G.S., Korablev N.M. (2014), “Time series forecasting on the basis of the case-based reasoning using the models of artificial immune systems”, System technologies, № 6(95), pp. 43-51, available at: http://nbuv.gov.ua/UJRN/st_2014_6_8
Ahmed N., Atiya A., Gayar N. (2010), “An Empirical Comparison of Machine Learning Models for Time Series Forecasting”, Econometric Reviews, № 29, pp. 594-621.
Makridakis S. (2000), “The M-3 Competition: Results, Conclusions and Implications”, International Journal of Forecasting, № 16, pp. 451-476.
Borghi P., Zakordonets O., Teixeira J. (2021), “A COVID-19 time series forecasting model based on MLP ANN”, Procedia Computer Science, № 181, pp. 940-947.
Namini S., Tavakoli N., Namin A. (2018), “A Comparison of ARIMA and LSTM in Forecasting Time Series”, 17th IEEE International Conference on Machine Learning and Applications.
Cao J., Li Z., Li J. (2019), “Financial time series forecasting model based on CEEMDAN and LSTM”, Physica A: Statistical Mechanics and its Applications, №. 519, pp. 127-139.
Vasily D., Matviychuk A., Datsenko N., Bezkorovainyi V., Azaryan A. (2020), “Machine learning approaches for financial time series forecasting”, Proceedings of the Selected Papers of the Special Edition of International Conference on Monitoring, Modeling & Management of Emergent Economy, № 2713, Odessa, Ukraine, pp. 434-450.
Hemmati A., Abdoos M., Akbar A. (2015), “Short term load forecasting using a hybrid intelligent method”, Knowledge-Based System, № 76, pp .139-147.
Mohammed A.S., Ivashchenko H., Filimonchuk T., Ivanisenko I., Barkovska O. (2020), "Green Hybrid Models Based on Clonal Selection and Case-based Reasoning for Short-term Time Series Forecasting", Journal of Green Engineering, Vol. 10, Issue 5. pp. 2139-2154, available at: http://www.jgenng.com/volume10-issue5.php
Hamzacebi C., Akay D., Kutay F. (2009), “Comparison of direct and iterative artificial neural network forecast approaches in multi-periodic time series forecasting”, Expert Systems with Applications, № 36, pp. 3839-3844.
Voyant C., Nivet M., Paoli C., Muselli M., Notton G. (2014), “Meteorological time series forecasting based on MLP modelling using heterogeneous transfer functions”, Journal of Physics: Conference Series, № 574, Madrid, Spain, pp. 28-31.
Greff, K., Srivastava, R., Koutnik, J., Steunebrink, B., Schmidhuber, J. (2017), “LSTM: A Search Space Odyssey”, IEEE Transactions on Neural Networks and Learning Systems, № 28, pp. 2222-2232.
Su Y., Kuo C. (2019), “On extended long short-term memory and dependent bidirectional recurrent neural network”, Neurocomputing, № 356, pp. 151-161.
