Сравнение современных методов прогнозирования доходностей криптовалют с бенчмарком SARIMA
Аннотация
Статья посвящена сравнительному анализу современных методов прогнозирования доходностей криптовалют. Рассматриваются классические статистические модели (SARIMA), методы машинного обучения, нейросетевые архитектуры (LSTM, трансформеры) и метод группового учета аргументов (МГУА). Цель исследования – выявить сильные и слабые стороны различных подходов при краткосрочном и среднесрочном прогнозировании.
Эмпирический анализ проведен на временных рядах 13 крупнейших криптовалют за период от 3 до 10 лет. Данные были преобразованы в логарифмические доходности, использовались блочная кросс-валидация и автоматизированный подбор гиперпараметров. Рассматривались горизонты 1, 7 и 30 дней, а также рекурсивные и масштабируемые методы многошагового прогнозирования.
Результаты показали, что для краткосрочных горизонтов наиболее точными оказались модели МГУА, тогда как LSTM демонстрировала худшие показатели. На месячном горизонте лидерство перешло к трансформеру Chronos, применявшемуся в режиме обучения внутри контекста. SARIMA сохранила устойчивость на средних горизонтах и подтвердила роль надежного бенчмарка. Масштабируемые методы позволили снизить ошибки по сравнению с рекурсивным прогнозом.
Полученные выводы подчеркивают отсутствие универсального алгоритма и важность выбора метода в зависимости от горизонта и задач анализа.
Скачивания
Литература
Aggarwal A., Gupta I., Garg N., Goel A. (2019) Deep Learning Approach to Determine the Impact of Socio Economic Factors on Bitcoin Price Prediction. 12th International Conference on Contemporary Computing (IC3), pp. 1–5.
Albariqi R., Winarko E. (2020) Prediction of Bitcoin Price Change using Neural Networks. International Conference on Smart Technology and Applications (ICoSTA), pp. 1–4.
Ansari A.F., Stella L., Turkmen C., Zhang X., Mercado P., Shen H., Shchur O., Rangapuram S.S., Arango S.P., Kapoor S., Zschiegner J., Maddix D.C., Wang H., Mahoney M.W., Torkkola K., Wilson A.G., Bohlke-Schneider M., Wang Y. (2024) Chronos: Learning the Language of Time Series. Transactions on Machine Learning Research.
Bayburin B., Mogilev P., Alexandrov M., Cardiff J., Koshulko O. (2020) Joint Mid-Term Forecast of Cryptocurrencies in Technique of Inductive Modelling (on Example of XRP, Waves, ETH). 31st Conference of FRUCT Association.
Box G.E.P., Jenkins G.M., Reinsel G.C., Ljung G.M. (2015) Time Series Analysis: Forecasting and Control. 5th ed. John Wiley and Sons Inc. Vol. 5.
Chen M., Narwal N., Schultz M. (2019) Predicting Price Changes in Ethereum. International Journal of Computation Science and Engineering.
Chen T., Guestrin C. (2016) XGBoost: A Scalable Tree Boosting System. The 22nd ACM SIGKDD International Conference, pp. 785–794.
Herremans D., Low K. (2022) Forecasting Bitcoin Volatility Spikes from Whale Transactions and CryptoQuant Data Using Synthesizer Transformer Models [Report].
Hochreiter S., Schmidhuber J. (1997) Long Short-Term Memory. Neural Computation, 9, 8, pp. 1735–1780.
Hou A.J., Wang W., Chen C.Y., Härdle W.K. (2020) Pricing Cryptocurrency Options. Journal of Financial Econometrics, 18, 2, pp. 250–279.
Hyndman R.J., Khandakar Y. (2008) Automatic Time Series Forecasting: The Forecast Package for R. Journal of Statistical Software, 27, 3, pp. 1–22.
Hyndman R.J., Koehler A.B. (2006) Another Look at Measures of Forecast Accuracy. International Journal of Forecasting, 22, 4, pp. 679–688.
Ivakhnenko A.G. (1968) The Group Method of Data Handling – a Rival of the Method of Stochastic Approximation. Soviet Automatic Control, 13, 3, pp. 43–55.
Ji S., Kim J., Im H. (2019) A Comparative Study of Bitcoin Price Prediction Using Deep Learning. Mathematics, 7, 10, p. 898.
John D., Binnewies S., Stantic B. (2024) Cryptocurrency Price Prediction Algorithms: A Survey and Future Directions [Report].
Khan F., Khan F., Shaikh P. (2023) Forecasting Returns Volatility of Cryptocurrency by Applying Various Deep Learning Algorithms. Future Business Journal, 9.
Kilimci H., Yildirim M., Kilimci Z. (2021) The Prediction of Short-Term Bitcoin Dollar Rate (BTC/USDT) using Deep and Hybrid Deep Learning Techniques. 5th International Symposium on Multidisciplinary Studies and Innovative Technologies (ISMSIT), pp. 633–637.
Kilimci Z. (2020) Sentiment Analysis Based Direction Prediction in Bitcoin Using Deep Learning Algorithms and Word Embedding Models. International Journal of Intelligent Systems and Applications, 8, pp. 60–65.
Kostin N., Belousov A. (2022) Implementation and Research of a Solution for Forecasting Crypto-currency Exchange Rate Dynamics Using Autoregressive Models. International Scientific Conference Proceedings “Advanced Information Technologies and Scientific Computing”, Publishing House of the Samara Scientific Center of the Russian Academy of Sciences, pp. 170–173 (In Russ.)
Mahfooz A., Phillips J. (2024) Conditional Forecasting of Bitcoin Prices Using Exogenous Variables. IEEE Access, 1, p. 1.
Marcellino M., Stock J.H., Watson M.W. (2006) A Comparison of Direct and Iterated Multistep AR Methods for Forecasting Macroeconomic Time Series. Journal of Econometrics, 135(1–2), pp. 499–526.
McNally S., Roche J., Caton S. (2018) Predicting the Price of Bitcoin Using Machine Learning. 26th Euromicro International Conference on Parallel, Distributed and Network-based Processing (PDP), pp. 339–343.
Pabuçcu H., Ongan S., Ongan A. (2023) Forecasting the Movements of Bitcoin Prices: An Application of Machine Learning Algorithms [Report].
Pečiulis T., Ahmad N., Menegaki A.N., Bibi A. (2024) Forecasting of Cryptocurrencies: Mapping Trends, Influential Sources, and Research Themes. Journal of Forecasting, 43.
Radford A., Jozefowicz R., Sutskever I. (2017) Learning to Generate Reviews and Discovering Sentiment.
Radford A., Narasimhan K., Salimans T., Sutskever I. (2018) Improving Language Understanding by Generative Pre-Training.
Ranco G., Aleksovski D., Caldarelli G., Grčar M., Mozetic I. (2015) The Effects of Twitter Sentiment on Stock Price Returns. PloS one, 10.
Seifaddini M., Habibdoust A. (2023) Improving Bitcoin Price Prediction Power by Time-Scale Decomposition and GMDH-Type Neural Network: A Comparison of Different Periods and Features. Journal of Mathematical Modeling (JMM), 11, 1.
Serafini G., Ping Y., Zhang Q., Brambilla M., Wang J., Hu Y., Li B. (2020) Sentiment-Driven Price Prediction of the Bitcoin based on Statistical and Deep Learning Approaches. International Joint Conference on Neural Networks (IJCNN), pp. 1–8.
Shamshad H., Ullah F., Ullah A., Kebande V.R., Ullah S., Al-Dhaqm A. (2023) Forecasting and Trading of the Stable Cryptocurrencies with Machine Learning and Deep Learning Algorithms for Market Conditions. IEEE Access 2023, 1, p. 99.
Shkanov B.A. (2024) Pricing Factors of Cryptocurrencies. Vestnik of Samara University. Economics and Management, 15, 1, pp. 225–237. (In Russ.)
Shkanov B.A. (2025) Comprehensive Approach to Portfolio Optimization Based on Modern Mathematical Methods. π-Economy,18 (2), pp. 179–201. (In Russ.)
Shkanov B., Alexandrov M. (2024) Social Influence, Market Manipulators, Hardware and Software As New Factors for Cryptocurrency Pricing: A Survey. Computación y Sistemas, 28 (3), pp. 1201–1207.
Singh S., Bhat M. (2024) Transformer-Based Approach for Ethereum Price Prediction Using Cross-currency Correlation and Sentiment Analysis [Report].
Sridhar S., Sanagavarapu S. (2021) Multi-Head Self-Attention Transformer for Dogecoin Price Prediction. 14th International Conference on Human System Interaction (HSI) At: Gdańsk, Poland.
Tan X., Kashef R. (2019) Predicting the Closing Price of Cryptocurrencies: A Comparative Study. E-Learning and Information Systems. DATA '19: Proceedings of the Second International Conference on Data Science, pp. 1–5.
Tandon S., Tripathi S., Saraswat P., Dabas C. (2019) Bitcoin Price Forecasting Using LSTM and 10-fold Cross Validation. International Conference on Signal Processing and Communication (ICSC).
Vaswani A., Shazeer N., Parmar N., Uszkoreit J., Jones L., Gomez A., Kaiser L., Polosukhin I. (2017) Attention Is All You Need.
Wu C.H., Lu C.C., Ma Y.F., Lu R.S. (2018) A New Forecasting Framework for Bitcoin Price with LSTM. IEEE International Conference on Data Mining Workshops (ICDMW).
