Applying artificial intelligence tools for time series analysis

N. Shapovalova І. Dotsenko A. Trachuk I. Skrynnikov
Received 22.12.2023, Revised 25.02.2024, Accepted 26.04.2024
Abstract

The purpose of the study is to analyze and compare the performance of artificial intelligence methods for time series forecasting problems. Research methods. The following research methods are used in the work: analysis of sources dealing with the subject under study, statistical methods, machine learning methods, neural networks, econometrics methods, optimization methods, methods for analyzing the interaction of external factors, methods for processing text information, formalization of constructed models, methods for designing software for development of a software model, empirical methods for substantiating the optimal parameters for training a model, methods of object-oriented design and programming The scientific novelty lies in the fact that the research can make a significant contribution to understanding and improving the effectiveness of artificial intelligence methods for analyzing time series, revealing their applicability in various fields and tasks; developing new algorithms for time series analysis based on artificial intelligence, which can take into account the specific requirements and characteristics of certain domains or data types, new approaches to the use and combination of various machine learning techniques to improve the accuracy and adaptability of models to real conditions. The practical value of the work done is that development and optimization of artificial intelligence methods can lead to improved accuracy of time series forecasting. This can have significant implications for manufacturing, finance, services, and other industries where accurate forecasts play an important role in decision-making. Using artificial intelligence for time series analysis can help in the efficient management of resources such as energy, manufacturing, transport, etc. Automated forecasting and optimization can help in more efficient use of resources and reduce costs. Results. The choice of effective artificial intelligence tools for forecasting various types of time series has been analyzed and justified, and the effectiveness of selected neural network architectures has been experimentally proven

Keywords:
time series; artificial intelligence methods; neural networks; forecasting

Retrieved from Vol. 22, No. 1, 2024

Pages 46–51

References Suggested citation

[1] Берзлев, О. Ю. (2013). Сучасний стан інформаційних систем прогнозування часових рядів. Управління розвитком складних систем, (13), 78-82.

[2] Kose, U., Arslan, A. (2018). Time series prediction with a hybrid system formed by artificial neural network and cognitive development optimization algorithm. Scientia Iranica, 26, 942-958. https://doi.org/10.24200/SCI.2018.20033.

[3] Wang, K. (2020). Artificial intelligence algorithm for optimal time series data model. IEEE Access. https://doi.org/10.1109/access.2020.2981488.

[4] Che, Z., Purushotham, S., Cho, K., Sontag, D., Liu, Y. (2016). Recurrent Neural Networks for Multivariate Time Series with Missing Values. Scientific Reports, 8. https://doi.org/10.1038/s41598-018-24271-9.

[5] Cinar, Y., Mirisaee, H., Goswami, P., Gaussier, É., Aït-Bachir, A. (2018). Period-aware content attention RNNs for time series forecasting with missing values. Neurocomputing, 312, 177- 186. https://doi.org/10.1016/j.neucom.2018.05.090.

[6] Lei, J., Liu, C., Jiang, D. (2019). Fault diagnosis of wind turbine based on Long Short-term memory networks. Renewable Energy. https://doi.org/10.1016/J.RENENE.2018.10.031.

[7] Kim, T., King, B. (2020). Time series prediction using deep echo state networks. Neural Computing and Applications, 1-19. https://doi.org/10.1007/s00521-020-04948-x.

[8] Єгоров, В. Ю. (2018). Алгоритми глибокого навчання у прогнозуваннi часових рядiв.

[9] Tolvanen, J., Kelly, S. (2010). Integrating models with domain-specific modeling languages. , 10:1- 10:6. https://doi.org/10.1145/2060329.2060354.

[10] Liang, H., Song, L., Wang, J., Guo, L., Li, X., Liang, J. (2021). Robust unsupervised anomaly detection via multi-time scale DCGANs with forgetting mechanism for industrial multivariate time series. Neurocomputing, 423, 444-462. https://doi.org/10.1016/j.neucom.2020.10.084.

[11] Schmidl, S., Wenig, P., Papenbrock, T. (2022). Anomaly detection in time series. Proceedings of the VLDB Endowment. https://doi.org/10.14778/3538598.3538602.

[12] Liu, S., Zhou, B., Ding, Q., Hooi, B., Zhang, Z., Shen, H., Cheng, X. (2023). Time Series Anomaly Detection With Adversarial Reconstruction Networks. IEEE Transactions on Knowledge and Data Engineering, 35, 4293-4306. https://doi.org/10.1109/TKDE.2021.3140058.

[13] Zhang, G. (2003). Time series forecasting using a hybrid ARIMA and neural network model. Neurocomputing, 50, 159-175. https://doi.org/10.1016/S0925-2312(01)00702-0.

[14] Сабат, А. О. (2020). Дослідження методів прогнозування часових рядів криптовалют. Архів кваліфікаційних робіт.

[15] Che, Z., Purushotham, S., Cho, K., Sontag, D., Liu, Y. (2016). Recurrent Neural Networks for Multivariate Time Series with Missing Values. Scientific Reports, 8. https://doi.org/10.1038/s41598-018-24271-9. 

Shapovalova, N., Dotsenko, І., Trachuk, A., & Skrynnikov, I. (2024). Applying artificial intelligence tools for time series analysis. Journal of Kryvyi Rih National University, 22(1), 46-51. https://doi.org/10.31721/2306-5451-2024-1-58-46-52
en