A Comparative Analysis of ARIMA and LSTM Models for Forecasting Non-Stationary Financial Time Series
Pages
1-11Keywords:
Abstract
Stock prices have become relatively wilder and nonlinear in nature over time. As such, outcome forecasting ability is essential for decision-making in finance. This research compares an adaptive Long Short-Term Memory neural net to the traditional ARIMA model in terms of predictability for a commonly used financial time series that is known to be non-stationary. Preprocessing using Normalizer and Augmented Dickey-Fuller (ADF) test for stationarity was carried out on Historical daily stock Close Price data collected from Yahoo Finance. Based on ACF/PACF analysis, an ARIMA (5,1,0) model was developed, while a multi-layer LSTM captured long-run dependencies. The Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and R-square (R2) were used to evaluate the models. The results revealed that LSTM outperformed ARIMA, with MSEs of 0.0023 and 0.0456, respectively. In addition, the LSTM model was more robust against sudden price variations, with an R2 value of 0.92 versus 0.857 for ARIMA. Such findings show that while ARIMA remains useful for detecting linear trends, adaptive deep learning models indicate that LSTM is far more effective in the case of dynamic, non-stationary environments. Future studies must thus explore hybrid architectures that take advantage of both approaches.
References
- Agarwal, H., Mahajan, G., Shrotriya, A., & Shekhawat, D. (2024). Predictive data analysis: Leveraging RNN and LSTM techniques for a time series dataset. Procedia Computer Science, 235, 979–989, January 2024. https://doi.org/10.1016/j.procs.2024.04.093
- Arik, S. O., Yoder, N. C., & Pfister, T. (2022). Self-adaptive forecasting for improved deep learning on non-stationary time-series. arXiv.org. Available: https://arxiv.org/abs/2202.02403
- Hu, J. (2024). A high-performance stock prediction system leveraging LSTM neural networks. Applied and Computational Engineering, 114(1), 65–72, December 2024. https://doi.org/10.54254/2755-2721/2024.18219
- Iaousse, M., Jouilil, Y., Bouincha, M., & Mentagui, D. (2023). Comparative simulation study of classical and machine learning techniques for forecasting time series data. International Journal of Online Engineering (iJOE), 19(8), 56–65. https://doi.org/10.3991/ijoe.v19i08.39853
- Kolambe, M., & Arora, S. (2024). Forecasting the future: A comprehensive review of time series prediction techniques. Journal of Electrical Systems, April 2024. https://doi.org/10.52783/jes.1478
- Li, Q., Fu, Y., Zhou, X., & Xu, Y. (2010). A hybrid support vector regression for time series prediction. Knowledge Discovery and Data Mining, 506–509. Available: https://doi.org/10.1109/WKDD.2010.92
- Melina, M., Sukono, S., Napitupulu, H., Mohamed, N., Chrisnanto, Y. H., Hadiana, A. I., et al. (2024). Comparative analysis of time series forecasting models using ARIMA and neural network autoregression methods. Barekeng, October 2024. https://doi.org/10.30598/barekengvol18iss4pp2563-2576
- Oancea, B., & Simionescu, M. (2024). Gross domestic product forecasting: Harnessing machine learning for accurate economic predictions in a univariate setting. Electronics, 13(24), 4918, December 2024. https://doi.org/10.3390/electronics13244918
- Pal, S., & Kar, S. (2021). Fuzzy transfer learning in time series forecasting for stock market prices. Research Square. Available: https://doi.org/10.21203/rs.3.rs-1015226/v1
- Pan, R. (2010). Holt–Winters exponential smoothing. In Wiley Encyclopedia of Operations Research and Management Science. Wiley. https://doi.org/10.1002/9780470400531.EORMS0385
- Rizvi, M. F. (2024). ARIMA model time series forecasting. International Journal for Research in Applied Science and Engineering Technology, May 2024. https://doi.org/10.22214/ijraset.2024.62416
- Rao, R. B., Rickard, S., & Coetzee, F. M. (1998). Time series forecasting from high-dimensional data with multiple adaptive layers. Knowledge Discovery and Data Mining, 319–324, August 1998. Available: https://www.aaai.org/Papers/KDD/1998/KDD98-057.pdf
- Sushanth, T., Siddarda, T. S., Sathvika, A., Shruthi, A. S. S., Lekha, A., & Kumar, T. S. (2024). Time series forecasting using RNN. Indian Scientific Journal of Research in Engineering and Management, 8(11), 1–8. https://doi.org/10.55041/ijsrem39164
- Shirley, C. P., Jingle, B. J., Abisha, M. B., Rajendran, V., & Absin, S. J. (2024). Reinforcement learning-based adaptive healthcare decision support systems using time series forecasting. In 2024 5th International Conference on Data Intelligence and Cognitive Informatics (ICDICI) (pp. 1476-1481). IEEE. https://doi.org/10.1109/icdici62993.2024.10810877
- Suddala, S. (2024). Dynamic demand forecasting in supply chains using hybrid ARIMA-LSTM architectures. International Journal of Advanced Research, 12(10), 1167–1171, October 2024. https://doi.org/10.21474/ijar01/19738
- Tulli, S. K. C. (2020). Comparative analysis of traditional and AI-based demand forecasting models. International Journal of Emerging Trends in Science and Technology, 6933–6956, June 2020. https://doi.org/10.18535/ijetst/v7i6.02
- Taslim, D. G., & Murwantara, I. M. (2024). Comparative analysis of ARIMA and LSTM for predicting fluctuating time series data. Buletin Teknik Elektro dan Informatika, June 2024. https://doi.org/10.11591/eei.v13i3.6034
- Teng, F. (2024). Beyond traditional forecasting: Machine learning and adaptive algorithms in EV sales predictions. Applied and Computational Engineering, 82(1), 24–28. https://doi.org/10.54254/2755-2721/82/2024glg0081
- Hamiane, S., Ghanou, Y., Khalifi, H., & Telmem, M. (2024). Comparative analysis of LSTM, ARIMA, and hybrid models for forecasting future GDP. Ingénierie des Systèmes d’Information, 29(3), 853–861, June 2024. https://doi.org/10.18280/isi.290306
