Сomparative analysis of machine learning algorithms for market capitalization time series forecasting
Article Sidebar
Main Article Content
Abstract
This paper presents a comparative analysis of four time series forecasting algorithms applied to market capitalization data of the world’s leading companies: Time Series Transformer, Recurrent Neural Network (RNN), Autoregressive Integrated Moving Average (ARIMA) and Simple Moving Average (SMA). The study is based on monthly market cap data for 1000 companies from 2000 to 2025, collected via Yahoo Finance. In addition to temporal dynamics, static categorical features such as sector, industry and market cap category were considered. The models were evaluated using MASE, sMAPE and MAPE metrics. Results show that the transformer-based model achieved the highest accuracy (MASE = 2.01, sMAPE = 15.63%, MAPE = 17.44%), confirming its suitability for long-term forecasting, especially when categorical features are incorporated. ARIMA and RNN showed moderate performance, while SMA performed the worst. Visualization further confirmed the transformer’s ability to capture seasonal patterns and trends. Future work includes integrating macroeconomic indicators to enhance prediction accuracy
Article Details
Issue
Section
This work is licensed under a Creative Commons Attribution 4.0 International License.
References
1. Ian Goodfellow, Yoshua Bengio, Aaron Courville, Deep Learning, MIT Press, 2016, 800 pp., ISBN 9780262035613.
2. Yue Chen, Mohan Li, Economic Forces in Stock Returns, arXiv preprint arXiv:2401.04132, 2024. https://arxiv.org/abs/2401.04132
3. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, Illia Polosukhin, Attention is all you need, Advances in Neural Information Processing Systems, vol. 30, 2017, pp. 5998–6008.
4. Bryan Lim, Sercan Ö. Arık, Nicolas Loeff, Tomas Pfister, Temporal Fusion Transformers for interpretable multi-horizon time series forecasting, International Journal of Forecasting, vol. 37, no. 4, 2021, pp. 1748–1764.
5. Angelo Ranaldo, Paul Söderlind, Safe haven currencies, Review of Finance, vol. 14, no. 3, 2010, pp. 385–407.
6. Ran Aroussi, yfinance: Python library for Yahoo Finance API, GitHub Repository, 2023. https://github.com/ranaroussi/yfinance
7. George E. P. Box, Gwilym M. Jenkins, Gregory C. Reinsel, Greta M. Ljung, Time Series Analysis: Forecasting and Control, 5th ed., Wiley, 2015, 712 pp., ISBN 9781118675021.
8. Benyamin Ghojogh, Recurrent Neural Networks and Long Short-Term Memory Networks: Tutorial and Survey, arXiv preprint arXiv:2304.11461, 2023. https://arxiv.org/abs/2304.11461
9. Hajar Gohari, Xuan-Hong Dang, Syed Shah, et al., Modality-aware Transformer for Financial Time Series Forecasting, Proceedings of the 5th ACM International Conference on AI in Finance, 2024, pp. 677–685. https://doi.org/10.1145/3677052.3698654
10. Liyilei Su, Xumin Zuo, Rui Li, et al., A systematic review for transformer-based long-term series forecasting, Artificial Intelligence Review, vol. 58, 2025, article 80. https://doi.org/10.1007/s10462-024-11044-2
11. Zhicheng Zhang, Yong Wang, Shaoqi Tan, et al., Enhancing Transformer-based models for long sequence time series forecasting via structured matrix, Neurocomputing, vol. 625, 2025, pp. 1–12.
12. Zhen Fang, Xu Ma, Huifeng Pan, Gonzalo R. Arce, Movement forecasting of financial time series based on adaptive LSTM-BN network, Expert Systems with Applications, vol. 213, March 2023, 119207. https://doi.org/10.1016/j.eswa.2022.119207
13. Diaa Salman, Cem Direkoglu, Mehmet Kusaf, Murat Fahrioglu, Hybrid deep learning models for time series forecasting of solar power, Neural Computing and Applications, vol. 36, no. 16, 2024. https://doi.org/10.1007/s00521-024-09558-5
14. Ying Zhao, Han Sun, Financial asset allocation strategies using statistical and machine learning models, Applied Soft Computing, vol. 150, 2024.
15. Vasyl Martsenyuk, Natalia Kit, A multivariate method of forecasting the nonlinear dynamics of production network based on multilayer neural models, Scientific Journal of TNTU (Ternopil), vol. 114, no. 2, 2024, pp. 39–50. https://doi.org/10.33108/visnyk_tntu2024.02.039