Prediction of the prices of digital cryptocurrencies in a high-frequency data world
Chenggang Li, Xuan Wang, Yan Gong, Liang Wu, Huiyang Li
Keywords:
Digital cryptocurrencies, Complementary Ensemble Empirical Mode Decomposition, Long Short-Term Memory Model, High-frequency data
Abstract:
This study predicts the prices of digital cryptocurrencies, which are non-linear, highly volatile, multi-scale, noisy, and, therefore, considerably more difficult to forecast compared to other financial products. Despite the proliferation of cryptocurrency markets, research on digital cryptocurrency prices remains scarce, while the existing research often fails to elucidate the characteristics of digital cryptocurrency prices. We developed CEEMD-LSTM—a hybrid prediction model based on the decomposition-reconstruction-integration framework—to predict digital cryptocurrency price movements. The CEEMD method is applied to decompose prices into high-frequency components, low-frequency components, and trend sequences, and reconstruct the combined IMF components. Thereafter, we apply the LSTM model to predict each component separately and obtain the sorted predicted values of prices by adding and integrating the predicted values of each component. Using four time-intervals of high-frequency data of Bitcoin and Ether, we conduct a systematic comparison of the prediction accuracy between CEEMD-LSTM and other models, single or combined. Results reveal that the prediction effect of the CEEMD-LSTM model consistently outperforms that of other models in the high-frequency data world. These findings enhance the understanding and prediction of digital cryptocurrency price movements. Additionally, they enable investors to manage risks and promote market competitiveness, offer reliable support for decision-making, address uncertainty arising from volatility, facilitate timely strategy adjustments to exploit opportunities and identify potential risks, and provide new methods for risk mitigation.
Fulltext download:
 Prediction of the prices of digital cryptocurrencies in a high-frequency data world [PDF file] [Filesize: 856.69 KB]
 10.7441/joc.2026.01.09
Li, C., Wang, X., Gong, Y., Wu, L., and Li, H. (2026). Prediction of the prices of digital cryptocurrencies in a high-frequency data world. Journal of Comptitiveness (18)1, 194-218. https://doi.org/10.7441/joc.2026.01.09
|
