BITCOIN PRICE PREDICTION WITH ARIMAX AND DEEP LEARNING MODELS

Abstract

Deep learning has emerged as a widely applied approach across various fields, with finance and forecasting being among its most prominent areas of use. Within this domain, different deep learning architectures have been developed to address specific prediction problems. This study compares the performance of ARIMAX and several deep learning models—including LSTM, BILSTM, CNN-LSTM, GRU, and TFT—in forecasting Bitcoin prices. The dataset consists of daily values from January 2014 to January 2025. The dependent variable is the daily Bitcoin closing price ($), while the independent variables include oil price (USD/barrel), gold price (USD/ounce), platinum price ($/XPT), and the USD/TRY exchange rate. All analyses were conducted in Python using Google Colab, with the Keras library employed for model implementation. Root Mean Square Error (RMSE) was selected as the evaluation metric for predictive accuracy. The results indicate that the TFT model achieved the highest predictive performance, followed closely by the GRU model. LSTM, BILSTM, and ARIMAX models showed similar yet weaker performance, while the CNN-LSTM model produced the least accurate forecasts, with significantly higher RMSE values compared to the other models.

Keywords

• Bitcoin
• Deep Learning
• ARIMAX
• Forecasting

References

1. [1] Akbulut S. Investigation of the relationship between BIST100 index and commodity, foreign exchange rates and stock indices of developing countries by using machine learning and deep learning methods [master’s thesis]. Aksaray, Turkey: Aksaray University; 2022
2. [2] Atlan F. Prediction of cryptocurrency values with artificial intelligence techniques [master’s thesis]. Burdur, Turkey: Mehmet Akif Ersoy University; 2019.
3. [3] Büyükkör Y. Bitcoin price prediction with deep learning and econometric model: LSTM and ARIMA. KMU J Soc Econ Res. 2024;26(47):978-993.
4. [4] Chung J, Gulcehre C, Cho K, Bengio Y. Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv. Published 2014. doi:10.48550/arXiv.1412.3555
5. [5] Demirci E, Karaatlı M. Forecasting cryptocurrency prices with LSTM and GRU models. J Mehmet Akif Ersoy Fac Econ Adm Sci. 2023;10(1):134-157.
6. [6] Farooq A, Uddin IM, Adnan M, Alarood AA, Alsolami E, Habibullah S. Interpretable multi-horizon time series forecasting of cryptocurrencies by leverage temporal fusion transformer. Heliyon. 2024;10:e40142.
7. [7] Gayhardt L, Garimella A, Salgado S, et al. Deep learning vs. machine learning in Azure Machine Learning. Microsoft Learn website. Published January 19, 2024. Accessed August 2, 2025. https://learn.microsoft.com/en-us/azure/machine-learning/concept-deep-learning-vs-machine-learning?view=azureml-api-2
8. [8] Guo Q, Lei S, Ye Q, Fang Z. MRC-LSTM: a hybrid approach of multi-scale residual CNN and LSTM to predict Bitcoin price. In: International Joint Conference on Neural Networks (IJCNN); May 2021.

9. [9] Güney EN. Stock price prediction with statistical methods in deep learning [master’s thesis]. Isparta, Turkey: Süleyman Demirel University Institute of Social Sciences; 2022.
10. [10] Karataş G. Deep learning based intrusion detection system [doctoral thesis]. Istanbul, Turkey: Marmara University Institute of Science and Technology; 2020.
11. [11] Köse N, Gür YE, Ünal E. Deep learning and machine learning insights into the global economic drivers of the Bitcoin price. J Forecast. 2025;0:1-33.
12. [12] Lim B, Arik SÖ, Loe N, Pfister T. Temporal fusion transformers for interpretable multi-horizon time series forecasting. Int J Forecast. 2021;37(4):1748-1764.
13. [13] Livieris IE, Pintelas E, Pintelas P. A CNN-LSTM model for gold price time-series forecasting. Neural Comput Appl. 2020;32:17351-17360.
14. [14] Lødøen S, Myklebust M. Predicting Bitcoin volatility using temporal fusion transformer [master’s thesis]. Norway: Norwegian University of Science and Technology, Department of Industrial Economics and Technology Management; 2024.
15. [15 McCulloch WS, Pitts W. A logical calculus of ideas immanent in nervous activity. Bull Math Biophys. 1943;5:115-133.
16. [16] Metin S. Bitcoin price analysis with deep learning methods. Munzur Univ J Soc Sci. 2025;3(2):93-110.
17. [17] Modi PD, Arshi K, Kunz PJ, Zoubir AM. A data-driven deep learning approach for Bitcoin price forecasting. arXiv. Published October 2023. doi:10.48550/arXiv.2311.06280
18. [18] Omole O, Enke D. Deep learning for Bitcoin price direction prediction: models and trading strategies empirically compared. Financ Innov. 2024;10:117.
19. [19] Öztürk K, Şahin ME. An overview of artificial neural networks and artificial intelligence. Takvim-i Vekayi. 2018;(32):25-36.
20. [20] Pankratz A. Forecasting with Dynamic Regression Models. Indiana: John Wiley & Sons; 1991.
21. [21] Polat OG. Bitcoin price forecasting with LSTM model [master’s thesis]. Istanbul, Turkey: Istanbul Commerce University; 2024.
22. [22] Şimşek Aİ. Predicting Bitcoin price: comparative analysis of machine learning and deep learning models. J Def Sci. 2024;20(2):327-342.
23. [23] Tanışman S, Karcıoğlu AA, Uğur A, Bulut H. Forecasting and comparison of Bitcoin price using LSTM network and ARIMA time series model. Eur J Sci Technol. 2021;(32):514-520.
24. [24] Varürer İ, Özaydın Ö. Silver price prediction with BILSTM deep learning model. In: Alpu Ö, ed. Current Topics in Statistics. Eğitim Publishing House; 2024:123-145