№1, 2020


Yadigar N. Imamverdiyev

This study proposes a deep neural network architecture for forecasting the Bitcoin exchange price, which is based on LSTM (Long-Short Term Memory) units, one of the types of recurrent neural networks. In recent years, cryptocurrencies have become important financial instruments, and the problem of forecasting cryptocurrency exchange rates for traditional currencies has become very urgent. Bitcoin is the first cryptocurrency, and more than half of the total cryptocurrency capitalization belongs to it and it plays the role of gold in the world of cryptocurrencies, that is, the price of other cryptocurrencies is often expressed in bitcoins. Based on these considerations, the problem of forecasting the Bitcoin exchange rate is considered, and the proposed deep LSTM architecture, in the experiments with real high-volume data spanning several years, shows better results than the statistical methods commonly used in forecasting of time series. The results obtained are important both for ordinary users interested in cryptocurrencies and for investors who are actively working in the field of cryptocurrencies. The results also show that deep learning approaches can be very productive when used in other tasks of the intellectual analysis of non-stationary time series on cryptocurrencies (pp.82-89).

Keywords: Bitcoin, cryptocurrency, time-series, forecasting, deep learning, recurrent neural networks, LSTM.
  • Narayanan A., Bonneau J., Felten E., Miller A., and Goldfeder S. Bitcoin and cryptocurrency technologies: A comprehensive introduction. Princeton University Press, 2016, 336 p.
  • Fry J., and Cheah E.T. Negative bubbles and shocks in cryptocurrency markets // International Review of Financial Analysis, 2016, vol.47, pp.343–352.
  • İmamverdiyev Y., Sadıyeva F. Kriptovalyuta verilənlərinin intelektual analizi məsələləri / “İnformasiya təhlükəsizliyinin aktual multidissiplinar elmi-praktiki problemləri” IV respublika konfransı, 14 dekabr 2018, s.47–52.
  • Catania L., Grassi S., and Ravazzolo F. Predicting the volatility of cryptocurrency time-series // Mathematical and Statistical Methods for Actuarial Sciences and Finance, 2018, pp.203–207.
  • Azari A. Bitcoin price prediction: An ARIMA approach. arXiv preprint arXiv:1904.05315. 2019.
  • Alessandretti L., ElBahrawy A., Aiello L.M., and Baronchelli A. Anticipating cryptocurrency prices using machine learning // Complexity, 2018, vol.2018, Article ID 8983590, 16 p. https://doi.org/10.1155/2018/8983590
  • Dingli A. and Fournier K. Financial time series forecasting – A deep learning approach // International Journal of Machine Learning and Computing, 2017, vol.7, pp.118–122.
  • Romeu P., Zamora-Martínez F., Botella-Rocamora P., Pardo J. Stacked denoising auto-encoders for short-term time series forecasting. In: Koprinkova-Hristova P., Mladenov V., Kasabov N. (eds) Artificial Neural Networks. Springer Series in Bio-/Neuroinformatics, Springer, Cham, 2015, vol.4. pp.463–486.
  • Bao W., Yue J., and Rao Y. A deep learning framework for financial time series using stacked autoencoders and long-short term memory // PloS One, 2017, vol.12, no.7, e0180944.
  • Nakamoto S. Bitcoin: A peer-to-peer electronic cash system. 2009. https://bitcoin.org/bitcoin.pdf.
  • Dyhrberg A.H. Bitcoin, gold and the dollar – A GARCH volatility analysis // Finance Research Letters, 2016, vol.16, pp.85–92.
  • Kristjanpoller W., and Minutolo M.C. A hybrid volatility forecasting framework integrating GARCH, artificial neural network, technical analysis and principal components analysis // Expert Systems with Applications, 2018, vol.109, pp.1–11.
  • Mallqui D.C., and Fernandes R.A. Predicting the direction, maximum, minimum and closing prices of daily Bitcoin exchange rate using machine learning techniques // Applied Soft Computing, 2019, vol.75, pp.596–606.
  • Lahmiri S., and Bekiros S. Cryptocurrency forecasting with deep learning chaotic neural networks // Chaos, Solitons & Fractals, 2019, vol.118, pp.35–40.
  • Pichl L., and Kaizoji T. Volatility analysis of Bitcoin price time series // Quantitative Finance and Economics, 2017, vol.1, no.4, pp.474–485. doi: 10.3934/QFE.2017.4.474.
  • Rebane J., Karlsson I., Denic S., and Papapetrou P. Seq2Seq RNNs and ARIMA models for cryptocurrency prediction: A comparative study / Proc. of SIGKDD Workshop on Fintech, 2018, pp.1–5.
  • McNally S., Roche J., and Caton S. Predicting the price of Bitcoin using Machine Learning / 26th IEEE Euromicro International Conference on Parallel, Distributed and Network-based Processing, 2018, pp.339–343.
  • Jang H., and Lee J. An Empirical Study on Modeling and Prediction of Bitcoin Prices with Bayesian Neural Networks Based on Blockchain Information // IEEE Access, 2017, vol.6, pp.5427–5437.
  • Madan I., Saluja S., and Zhao A. Automated bitcoin trading via machine learning algorithms, 2015. Dept. Comput. Sci., Stanford Univ., Stanford, CA, USA, Tech. Rep., 
    URL: http://cs229. stanford. edu/proj2014/Isaac% 20Madan, 20.
  • Indera N. I., Yassin I. M., Zabidi A., and Rizman Z. I. Non-linear autoregressive with exogeneous input (NARX) Bitcoin price prediction model using PSO-optimized parameters and moving average technical indicators // Journal of Fundamental and Applied Sciences, 2017, vol.9, no.3S, pp.791–808.
  • Sin E., and Wang L. Bitcoin price prediction using ensembles of neural networks / 13th IEEE International Conference on Natural Computation, Fuzzy Systems and Knowledge Discovery, 2017, pp.666–671.
  • Karalevicius V., Degrande N., & De Weerdt J. Using sentiment analysis to predict interday Bitcoin price movements // The Journal of Risk Finance, 2018, vol.19, no.1, pp.56–75.
  • Kodama O., Pichl L., and Kaizoji T. Regime change and trend prediction for Bitcoin time series data / CBU International Conference Proceedings, 2017, vol.5, pp.384–388.
  • Hochreiter S., and Schmidhuber J. Long short-term memory // Neural computation, 1997, vol.9(8), pp.1735–1780.
  • Srivastava N., Hinton G., Krizhevsky A., Sutskever I., and Salakhutdinov R. Dropout: a simple way to prevent neural networks from overfitting // The Journal of Machine Learning Research, 2014, vol.15, no.1, pp.1929–1958.