Bitcoin Fiyat Tahmini: Makine Öğrenmesi ve Derin Öğrenme Yöntemlerine İlişkin Karşılaştırmalı Bir Analiz

Abstract

Bitcoin son dönemde bir yatırım aracı olarak görülmekte ve Bitcoine olan ilgi giderek artmaktadır. Bir finansal değerin fiyatının doğru tahmin edilmesi hem bireysel hem de kurumsal yatırımcılar açısından çok önemlidir. Bu çalışmanın amacı Bitcoin fiyatının tahmin edilmeside SVR, CNN, LSTM, CNN-BiLSTM ve XGBoost modellerinin hangisinin daha iyi performans gösterdiğinin belirlenmesi ve Bitcoin fiyatının tahmin edilmesi için en iyi algoritmanın ortaya çıkarılmasıdır. Bu çalışmada Bitcoin fiyatının tahmin edilmesi için S&P500 endeksi, Altın/Dolar kuru, West Texas Spot Petrol Fiyatı ve Dolar Endeksi kullanılmıştır. Veri seti 01.01.2015-18.09.2023 arasındaki 2191 günü kapsamaktadır. Çalışmada önerilen modeller önce eğitilmiş sonrasında test edilmiştir. Modellerin performanslarını değerlendirmek için RMSE, MAE, MAPE ve R2 istatistik katsayıları kullanılmıştır. Elde edilen sonuçlara göre Bitcoin fiyatının tahmin edilmesinde en iyi performansı bütün performans ölçütlerinde XGBoost algoritması vermiştir. XGBoost modelinden sonra en iyi sonuçları sırasıyla hibrit yaklaşım olan CNN-BiLSTM, CNN, LSTM modelleri vermiştir. En kötü performans ise SVR modeli tarafından üretilmiştir.

Keywords

• Zaman Serisi
• Bitcoin
• Derin Öğrenme
• Makine Öğrenmesi
• Fiyat Tahmini

Predicting Bitcoin Price: Comparative Analysis of Machine Learning and Deep Learning Models

Abstract

Bitcoin has become a prominent financial instrument in recent years, attracting increasing attention as a digital currency. Accurately forecasting the valuation of a financial asset carries substantial significance for both retail and institutional investors. The aim of this study is to evaluate and compare the predictive capabilities of various models, namely Support Vector Regression (SVR), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), a hybrid model combining CNN and Bidirectional LSTM (CNN-BiLSTM), and XGBoost, in the context of forecasting Bitcoin price. The main aim of this study is to ascertain the algorithm that demonstrates the most efficacy in forecasting the price of Bitcoin. This study utilizes the S&P500 index, Gold/Dollar exchange rate, West Texas Spot Oil Price, and Dollar Index as exogenous factors in order to forecast the price of Bitcoin. The dataset encompasses a consecutive time span of 2191 days, commencing on January 1, 2015 and concluding on September 18, 2023. The models outlined in the study undergo a two-stage procedure, including of training and testing. The assessment of the models' performance was carried out by utilizing several statistical measures, such as Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and R-squared (R2). The results indicate that the XGBoost algorithm had greater performance in projecting the price of Bitcoin, as evidenced by its consistently higher performance metrics across all evaluated aspects. The XGBoost model was succeeded by the CNN-BiLSTM, CNN, and LSTM models, which are hybrid methodologies, resulting in the most advantageous results. The SVR model demonstrated the least favorable performance..

Keywords

• Bitcoin
• Time series
• Deep Learning
• Machine Learning
• Price Prediction

References

1. [1] S. Nakamoto, “Bridging The Global Digital Divide Through Digital Inclusion: The Role Of ICT Access And ICT Use,” Transform. Gov. People, Process Policy, pp. 1–9, 2008.
2. [2] R. Murugesan, V. Shanmugaraja, and A. Vadivel, “Forecasting Bitcoin Price Using Interval Graph and ANN Model: A Novel Approach,” SN Comput. Sci., vol. 3, no. 5, pp. 1– 10, 2022, doi: 10.1007/s42979-022-01291-x.
3. [3] W. Yiying and Z. Yeze, “Cryptocurrency Price Analysis with Artificial Intelligence,” 5th Int. Conf. Inf. Manag. ICIM 2019, pp. 97–101, 2019, doi: 10.1109/INFOMAN.2019.8714700.
4. [4] Ferdiansyah, S. H. Othman, R. Zahilah Raja Md Radzi, D. Stiawan, Y. Sazaki, and U. Ependi, “A LSTM-Method for Bitcoin Price Prediction: A Case Study Yahoo Finance Stock Market,” ICECOS 2019 - 3rd Int. Conf. Electr. Eng. Comput. Sci. Proceeding, pp. 206–210, 2019, doi: 10.1109/ICECOS47637.2019.8984499.
5. [5] M. J. Hamayel and A. Y. Owda, “A Novel Cryptocurrency Price Prediction Model Using GRU, LSTM and bi-LSTM Machine Learning Algorithms,” Ai, vol. 2, no. 4, pp. 477–496, 2021, doi: 10.3390/ai2040030.
6. [6] S. Dong, “Virtual Currency Price Prediction Based on Segmented İntegrated Learning,” 2022 IEEE 2nd Int. Conf. Power, Electron. Comput. Appl., no. January, pp. 549–552, 2022, doi: 10.1109/icpeca53709.2022.9719070.
7. [7] A. Aggarwal, I. Gupta, N. Garg, and A. Goel, “Deep Learning Approach to Determine the Impact of Socio Economic Factors on Bitcoin Price Prediction,” 2019 12th Int. Conf. Contemp. Comput. IC3 2019, pp. 1–5, 2019, doi: 10.1109/IC3.2019.8844928.
8. [8] S. Cavalli and M. Amoretti, “CNN-Based Multivariate Data Analysis for Bitcoin Trend Prediction,” Appl. Soft Comput., vol. 101, p. 107065, 2021, doi: 10.1016/j.asoc.2020.107065.

9. [9] S. H. Hasan, S. H. Hasan, M. S. Ahmed, and S. H. Hasan, “A Novel Cryptocurrency Prediction Method Using Optimum CNN,” Comput. Mater. Contin., vol. 71, no. 1, pp. 1051–1063, 2022, doi: 10.32604/cmc.2022.020823.
10. [10] E. Edgari, J. Thiojaya, and N. N. Qomariyah, “The Impact of Twitter Sentiment Analysis on Bitcoin Price during COVID-19 with XGBoost,” 5th Int. Conf. Comput. Informatics, ICCI 2022, pp. 337–342, 2022, doi: 10.1109/ICCI54321.2022.9756123.
11. [11] R. G. Tiwari, A. K. Agarwal, R. K. Kaushal, and N. Kumar, “Prophetic Analysis of Bitcoin price using Machine Learning Approaches,” Proc. IEEE Int. Conf. Signal Process. Control, vol. 2021-Octob, pp. 428–432, 2021, doi: 10.1109/ISPCC53510.2021.9609419.
12. [12] Q. Guo, S. Lei, Q. Ye, and Z. Fang, “MRCLSTM: A Hybrid Approach of Multi-scale Residual CNN and LSTM to Predict Bitcoin Price,” Proc. Int. Jt. Conf. Neural Networks, vol. 2021-July, pp. 1–8, 2021, doi: 10.1109/IJCNN52387.2021.9534453.
13. [13] S. Kazeminia, H. Sajedi, and M. Arjmand, “Real-Time Bitcoin Price Prediction Using Hybrid 2D-CNN LSTM Model,” 2023 9th Int. Conf. Web Res. ICWR 2023, pp. 173–178, 2023, doi: 10.1109/ICWR57742.2023.10139275.
14. [14] Y. Li and W. Dai, “Bitcoin Price Forecasting Method Based on CNN‐LSTM Hybrid Neural Network Model,” J. Eng., vol. 2020, no. 13, pp. 344–347, 2020, doi: 10.1049/joe.2019.1203.
15. [15] I. E. Livieris, N. Kiriakidou, S. Stavroyiannis, and P. Pintelas, “An Advanced CNN-LSTM Model for Cryptocurrency Forecasting,” Electron., vol. 10, no. 3, pp. 1–16, 2021, doi: 10.3390/electronics10030287.
16. [16] P. C. Sekhar, M. Padmaja, B. Sarangi, and Aditya, “Prediction of Cryptocurrency using LSTM and XGBoost,” 2022 IEEE Int. Conf. Blockchain Distrib. Syst. Secur. ICBDS 2022, pp. 1–5, 2022, doi: 10.1109/ICBDS53701.2022.9935871.
17. [17] J. Chen, “Analysis of Bitcoin Price Prediction Using Machine Learning,” J. Risk Financ. Manag., vol. 16, no. 1, 2023, doi: 10.3390/jrfm16010051.
18. [18] D. Y. Lee and S. Y. Park, “Global Energy İntensity Convergence Using a Spatial Panel Growth Model,” Appl. Econ., vol. 00, no. 00, pp. 1–20, 2022, doi: 10.1080/00036846.2022.2131715.
19. [19] R. G. Tiwari, A. K. Agarwal, R. K. Kaushal, and N. Kumar, “Prophetic Analysis of Bitcoin Price Using Machine Learning Approaches,” Proc. IEEE Int. Conf. Signal Process. Control, vol. 2021-October, pp. 428–432, 2021, doi: 10.1109/ISPCC53510.2021.9609419.
20. [20] S. Erfanian, Y. Zhou, A. Razzaq, A. Abbas, A. A. Safeer, and T. Li, “Predicting Bitcoin (BTC) Price in the Context of Economic Theories: A Machine Learning Approach,” Entropy, vol. 24, no. 10, pp. 1–29, 2022, doi: 10.3390/e24101487.
21. [21] K. Sathiyapriya, S. Vankadara, K. S. Babu, and M. Muralidharan, “Performance Comparison of LSTM and XGBOOST for Ether Price Prediction from Spam Filtered Tweets,” Proc. 2023 Int. Conf. Intell. Syst. Commun. IoT Secur. ICISCoIS 2023, pp. 650–655, 2023, doi: 10.1109/ICISCoIS56541.2023.10100425.
22. [22] F. Sherratt, A. Plummer, and P. Iravani, “Understanding Lstm Network Behaviour Of İmu-Based Locomotion Mode Recognition For Applications İn Prostheses And Wearables,” Sensors (Switzerland), vol. 21, no. 4, pp. 1–23, 2021, doi: 10.3390/s21041264.
23. [23] S. Mehtab, J. Sen, and S. Dasgupta, “Robust Analysis of Stock Price Time Series Using CNN and LSTM-Based Deep Learning Models,” Proc. 4th Int. Conf. Electron. Commun. Aerosp. Technol. ICECA 2020, pp. 1481–1486, 2020, doi: 10.1109/ICECA49313.2020.9297652.
24. [24] A. Boru İpek, “Prediction of market-clearing price using neural networks based methods and boosting algorithms,” Int. Adv. Res. Eng. J., vol. 5, no. 2, pp. 240–246, 2021, doi: 10.35860/iarej.824168.
25. [25] H. Abar, “Altın Fiyatlarının Kestirimi,” vol. 83, no. Yaz, 2020.
26. [26] Y. Zouzou And H. Çıtakoğlu, “Reference Evapotranspiration Prediction From Limited Climatic Variables Using Support Vector Machines and Gaussian Processes,” Eur. J. Sci. Technol., no. 28, pp. 346–351, 2021, doi: 10.31590/ejosat.999319.
27. [27] H. Wang, J. Wang, L. Cao, Y. Li, Q. Sun, and J. Wang, “A Stock Closing Price Prediction Model Based on CNN-BiSLSTM,” Complexity, vol. 2021, 2021, doi: 10.1155/2021/5360828.
28. [28] LeCun, Y., Boser, B. E., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W. E., & Jackel, L. D., 1990. Handwritten Digit Recognition With A Backpropagation Network. In Advances in Neural Information Processing systems, 396-404.
29. [29] Cortes, C. & Vapnik, V. (1995). Support vector networks. Machine Learning, 20, 273–297.
30. [30] Vapnik, V. (1999). The nature of statistical learning theory. Springer science & business media.