WT-ConvLSTM Time Series Forecasting for Non-Stationary Characteristics of Cryptocurrency Prices
Abstract
coupling", this paper proposes a WT-ConvLSTM forecasting model that integrates Wavelet Transform (WT) and Convolutional Long ShortTerm Memory Network (ConvLSTM). By weakening non-stationary interference through WT decomposition and enhancing feature capture
capability via ConvLSTM, the model provides a new method with high accuracy and strong robustness for cryptocurrency price forecasting.
Keywords
Full Text:
PDFReferences
[1] Yang, M., Feng, H. L., Wang, X., et al. (2024). A literature review on the data element market: Value, pricing, and trading[J]. Chinese
Journal of Network and Information Security, 10(03), 1-19.
[2] Sorkin, M., Xiong, W., & Yan, C. F. (2024). Achieving decentralization through tokenization[J]. Digital Economy Development and
Governance, (01), 85-139. https://doi.org/CNKI: SUN: SZZL.0.2024-01-007
[3] Li, S. J. (2024). Analysis of systemic risk spillover effects and influencing factors of financial institutions (Master's Thesis)[D]. Shandong University, Jinan. https://doi.org/10.27272/d.cnki.gshdu.2024.004503
[4] Liu, R. Y. (2021). Multifractal analysis of stock, foreign exchange, and agricultural futures markets (Master's Thesis)[D]. Nanjing University of Finance and Economics, Nanjing. https://doi.org/10.27705/d.cnki.gnjcj.2021.000062
[5] Yang, K., & Yang, L. L. (2024). Empirical mode decomposition: Theoretical development and financial empirical progress[J]. Finance
Theory and Teaching, 42(04), 45-53. https://doi.org/10.13298/j.cnki.ftat.2024.04.011
DOI: http://dx.doi.org/10.70711/aitr.v3i3.8043
Refbacks
- There are currently no refbacks.
