Derin Öğrenme Modelleri ile Bitcoin Fiyatlarının Tahmini

Abstract

Bu çalışma, Bitcoin fiyatlarının bir adım ileri tahmini için dört derin öğrenme mimarisinin—GRU, LSTM, RNN ve CNN—tahmin performansını karşılaştırmaktadır. Optimum yapılandırma grid search yöntemiyle belirlenmiş ve adil bir değerlendirme için tüm modellere aynı şekilde uygulanmıştır. Ocak 2018–Temmuz 2025 dönemine ait günlük Bitcoin kapanış fiyatları kullanılarak yapılan analizde, GRU modelinin en düşük tahmin hatalarını (MSE, RMSE, MAE, MAPE) ve en yüksek R² değerini sağladığı, LSTM’in ise ona yakın bir performans gösterdiği bulunmuştur. Görsel analizler, özellikle dalgalı dönemlerde GRU ve LSTM’in gerçek fiyatlarla daha uyumlu seyrettiğini doğrulamaktadır. Ekonomik değer analizinde, model tahminleri %0,10 işlem maliyeti ve %0,10 işlem eşiği dikkate alınarak, kısa satışa izin verilen ve yalnızca uzun pozisyon içeren iki varyantla kural tabanlı alım-satım stratejisine entegre edilmiştir. Kısa satışlı GRU stratejisi, en yüksek nihai serveti (Al ve Tut stratejisinden yaklaşık %24 daha yüksek) ve yıllık bileşik büyüme oranı (CAGR), maksimum düşüş (Maximum Drawdown) ile Sharpe oranı üzerinden ölçülen daha üstün risk düzeltilmiş getirileri sağlamıştır. Bulgular, dikkatli hiperparametre optimizasyonu ile karmaşık zamansal bağımlılıkları yakalayabilen bir mimarinin kripto para piyasalarında hem tahmin doğruluğunu hem de alım-satım kârlılığını artırabileceğini göstermektedir. Bu sonuçlar, yapay zekâ destekli alım-satım sistemleri için pratik çıkarımlar sunmaktadır.

Keywords

• Bitcoin
• Derin öğrenme
• Sinir ağları
• Tahminleme

Forecasting Bitcoin Prices with Deep Learning Models

Abstract

This study compares the forecasting performance of four deep learning architectures—GRU, LSTM, RNN, and CNN—for one-step-ahead Bitcoin price prediction. A grid search determined the optimal configuration, which was applied uniformly across models to ensure fair evaluation. Using daily BTC closing prices from January 2018 to July 2025, it is found that the GRU model achieved the lowest forecasting errors (MSE, RMSE, MAE, MAPE) and the highest R², with LSTM performing closely behind. Visual analyses confirmed that GRU and LSTM maintained stronger alignment with actual prices during volatile periods. To assess economic value, model forecasts were integrated into a rule-based trading strategy under realistic market frictions, including a 0.10% transaction cost and a 0.10% trading threshold, with both short-selling-enabled and long-only variants tested. The GRU strategy with short-selling generated the highest terminal wealth (approximately 24% higher than the Buy-and-Hold benchmark) and superior risk-adjusted returns, measured by CAGR, Maximum Drawdown, and Sharpe Ratio. The findings demonstrate that careful hyperparameter optimization, coupled with an architecture capable of capturing complex temporal dependencies, can significantly improve both predictive accuracy and trading profitability in cryptocurrency markets. These results provide practical implications for designing AI-driven trading systems.

Keywords

• Bitcoin
• Deep learning
• Neural networks
• Forecasting

References

1. Aggarwal, A., Gupta, I., Garg, N., & Goel, A. (2019). Deep learning approach to determine the impact of socio-economic factors on Bitcoin price prediction. 2019 Twelfth International Conference on Contemporary Computing (IC3), 1–5. https://doi.org/10.1109/IC3.2019.8844928
2. Al-Zakhali, O. A., & Abdulazeez, A. M. (2024). Comparative analysis of machine learning and deep learning models for bitcoin price prediction. Indonesian Journal of Computer Science, 13(1). https://doi.org/10.33022/ijcs.v13i1.3722
3. Amirshahi, B., & Lahmiri, S. (2023). Investigating the effectiveness of Twitter sentiment in cryptocurrency close price prediction by using deep learning. Expert Systems, 42(1), 1-23. https://doi.org/10.1111/exsy.13428
4. Baur, D. G., Hong, K., & Lee, A. D. (2018). Bitcoin: Medium of exchange or speculative assets? Journal of International Financial Markets, Institutions and Money, 54, 177–189. https://doi.org/10.1016/j.intfin.2017.12.004
5. Böhme, R., Christin, N., Edelman, B., & Moore, T. (2015). Bitcoin: Economics, technology, and governance. Journal of Economic Perspectives, 29(2), 213–238. https://doi.org/10.1257/jep.29.2.213
6. Cho, K., van Merriënboer, B., Gulcehre, C., Bahdanau, D., Bougares, F., Schwenk, H., & Bengio, Y. (2014). Learning phrase representations using RNN encoder-decoder for statistical machine translation. arXiv:1406.1078. https://doi.org/10.48550/arXiv.1406.1078
7. Corbet, S., Cumming, D., Lucey, B., Peat, M., & Vigne, S. (2020). The destabilising effects of cryptocurrency cybercriminality. Economics Letters, 191, 108741. https://doi.org/10.1016/j.econlet.2019.108741
8. Dutta, A., Kumar, S. S., & Basu, M. (2020). A gated recurrent unit approach to bitcoin price prediction. Journal of Risk and Financial Management, 13(2), 23. https://doi.org/10.3390/jrfm13020023

9. El-Berawi, A. S., Belal, M. A. F., & Abd Ellatif, M. M. (2021). Adaptive deep learning based cryptocurrency price fluctuation classification. International Journal of Advanced Computer Science and Applications, 12(12), 64-72. https://doi.org/10.14569/IJACSA.2021.0121264
10. Elman, J. L. (1990). Finding structure in time. Cognitive Science, 14(2), 179–211. https://doi.org/10.1207/s15516709cog1402_1
11. Entrop, O., Frijns, B., & Seruset, M. (2020). The determinants of price discovery on Bitcoin markets. Journal of Futures Markets, 40(5), 816–837. https://doi.org/10.1002/fut.22101
12. Gholamy, A., Kreinovich, V., & Kosheleva, O. (2018). Why 70/30 or 80/20 relation between training and testing sets: A pedagogical explanation (Departmental Technical Report No. CS-2018-1209). University of Texas at El Paso. https://www.cs.utep.edu/vladik/2018/tr18-09.pdf
13. Hadi, M. H. A., Ramli, N. A., & Islam, Q. U. (2023). Predicting Bitcoin and Ethereum prices using long short-term memory and gated recurrent unit. Data Analytics and Applied Mathematics (DAAM), 8–17. https://doi.org/10.15282/daam.v4i2.10195
14. Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735–1780. https://doi.org/10.1162/neco.1997.9.8.1735
15. Karakoyun, E. Ş., & Çıbıkdiken, A. O. (2018). Comparison of ARIMA time series model and LSTM deep learning algorithm for Bitcoin price forecasting. The 13th Multidisciplinary Academic Conference in Prague 2018, 171–180.
16. Katsiampa, P. (2019). Volatility co-movement between Bitcoin and Ether. Finance Research Letters, 30, 221–227. https://doi.org/10.1016/j.frl.2018.10.005
17. Kaur, R., Uppal, M., Gupta, D., Juneja, S., Arafat, S. Y., Rashid, J., & Alroobaea, R. (2025). Development of a cryptocurrency price prediction model: Leveraging GRU and LSTM for Bitcoin, Litecoin and Ethereum. PeerJ Computer Science, 11, e2675. https://doi.org/10.7717/peerj-cs.2675
18. Lahmiri, S., & Bekiros, S. (2019). Cryptocurrency forecasting with deep learning chaotic neural networks. Chaos, Solitons & Fractals, 118, 35–40. https://doi.org/10.1016/j.chaos.2018.11.014
19. LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278–2324. https://doi.org/10.1109/5.726791
20. Li, L., Arab, A., Liu, J., Liu, J., & Han, Z. (2019). Bitcoin options pricing using LSTM-based prediction model and blockchain statistics. 2019 IEEE International Conference on Blockchain (Blockchain), 67–74. https://doi.org/10.1109/Blockchain.2019.00018
21. Lim, B., & Zohren, S. (2021). Time-series forecasting with deep learning: A survey. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 379(2194), 20200209. https://doi.org/10.1098/rsta.2020.0209
22. McNally, S., Roche, J., & Caton, S. (2018). Predicting the price of Bitcoin using machine learning. 2018 26th Euromicro International Conference on Parallel, Distributed and Network-Based Processing (PDP), 339–343. https://doi.org/10.1109/PDP2018.2018.00060
23. Nakamoto, S. (2008, 31 October). Bitcoin: A peer-to-peer electronic cash system. https://bitcoin.org/bitcoin.pdf
24. Phaladisailoed, T., & Numnonda, T. (2018). Machine learning models comparison for Bitcoin price prediction. 2018 10th International Conference on Information Technology and Electrical Engineering (ICITEE), 506–511. https://doi.org/10.1109/ICITEED.2018.8534911
25. Pichl, L., & Kaizoji, T. (2017). Volatility analysis of Bitcoin price time series. Quantitative Finance and Economics, 1(4), 474–485. https://doi.org/10.3934/QFE.2017.4.474
26. Pratas, T. E., Ramos, F., & Rubio, L. (2023). Forecasting Bitcoin volatility: Exploring the potential of deep learning. Eurasian Economic Review, 13(2), 285–305. https://doi.org/10.1007/s40822-023-00232-0
27. Shen, Z., Wan, Q., & Leatham, D. J. (2021). Bitcoin return volatility forecasting: A comparative study between GARCH and RNN. Journal of Risk and Financial Management, 14(7), 337. https://doi.org/10.3390/jrfm14070337
28. Siami-Namini, S., & Namin, A. S. (2018). Forecasting economics and financial time series: ARIMA vs. LSTM. arXiv:1803.06386. https://doi.org/10.48550/arXiv.1803.06386
29. Sifat, I. M., Mohamad, A., & Shariff, M. S. B. M. (2019). Lead-lag relationship between Bitcoin and Ethereum: Evidence from hourly and daily data. Research in International Business and Finance, 50, 306–321. https://doi.org/10.1016/j.ribaf.2019.06.012
30. Urquhart, A. (2017). Price clustering in Bitcoin. Economics Letters, 159, 145-148. https://doi.org/10.1016/j.econlet.2017.07.035
31. Wardak, A. B., & Rasheed, J. (2022). Bitcoin cryptocurrency price prediction using long short-term memory recurrent neural network. European Journal of Science and Technology, 38, 47–53. https://doi.org/10.31590/ejosat.1079622