|
|
基于多步分解的股价预测模型
|
Abstract:
金融领域一直备受关注，且股价受多种因素影响，其预测也存在一定的挑战性。为准确预测股价，为投资者和交易者提供有益的决策支持，本研究旨在提出一种基于多步分解的股价预测模型。首先通过变分模态分解(VMD)分解原始序列，重构高复杂成分，然后应用鲁棒局部均值分解(RLMD)进行二次分解，最后利用PSO-LSTM模型进行预测。为验证所提模型的有效性，将股票数据经VMD-MFE-RLMD分解与没有分解、只有VMD分解以及传统模型CNN、SVR、GRU进行对比，在沪深300指数数据集上的结果显示：股票数据经VMD-MFE-RLMD分解的预测误差MAE、MSE、RMSE、MAPE均小于没有分解以及只经VMD分解的预测误差，且低于传统预测模型的预测误差，提高了预测精度。最后，将此模型应用在上证50指数数据集上，同样取得了较好的预测结果，再次证明了所提模型在股价预测上具有更高的预测精度。
The financial sector has always been highly regarded, and stock prices are influenced by various factors, making their predictions somewhat challenging. To accurately predict stock prices and provide useful decision support for investors and traders, this study aims to propose a stock price prediction model based on multi-step decomposition. Firstly, the original sequence is decomposed using Variational Mode Decomposition (VMD) and the high complexity components are reconstructed. Then, robust Local Mean Decomposition (RLMD) is applied for secondary decomposition, and finally, the PSO-LSTM model is used for prediction. To verify the effectiveness of the proposed model, the stock data was decomposed by VMD-MFE-RLMD and compared with non-decomposition, only VMD decomposition, and traditional models such as CNN, SVR, and GRU. The results on the Shanghai and Shenzhen 300 Index dataset showed that the prediction errors MAE, MSE, RMSE, and MAPE of stock data decomposed by VMD-MFE-RLMD were smaller than those without decomposition and only VMD decomposition, and lower than those of traditional prediction models, improving prediction accuracy. Finally, applying this model to the Shanghai Stock Exchange 50 Index dataset also achieved good prediction results, once again proving that the proposed model has higher prediction accuracy in stock price prediction.
| [1] | Song, Y.P., et al. (2024) Multi-Decomposition in Deep Learning Models for Futures Price Prediction. Expert Systems with Applications, 246, Article ID: 123171. https://doi.org/10.1016/j.eswa.2024.123171 |
| [2] | Wang, J., Cui, Q., Sun, X. and He, M. (2022) Asian Stock Markets Closing Index Forecast Based on Secondary Decomposition, Multi-Factor Analysis and Attention-Based LSTM Model. Engineering Applications of Artificial Intelligence, 113, Article ID: 104908. https://doi.org/10.1016/j.engappai.2022.104908 |
| [3] | Mao, Y. and Yu, X. (2024) A Hybrid Forecasting Approach for China’s National Carbon Emission Allowance Prices with Balanced Accuracy and Interpretability. Journal of Environmental Management, 351, Article ID: 119873. https://doi.org/10.1016/j.jenvman.2023.119873 |
| [4] | Liu, Z., Jin, Y., Zuo, M.J. and Feng, Z. (2017) Time-Frequency Representation Based on Robust Local Mean Decomposition for Multicomponent AM-FM Signal Analysis. Mechanical Systems and Signal Processing, 95, 468-487. https://doi.org/10.1016/j.ymssp.2017.03.035 |
| [5] | Wang, H., Xiong, M., Chen, H. and Liu, S. (2022) Multi-Step Ahead Wind Speed Prediction Based on a Two-Step Decomposition Technique and Prediction Model Parameter Optimization. Energy Reports, 8, 6086-6100. https://doi.org/10.1016/j.egyr.2022.04.045 |
| [6] | Ngoc-Lan Huynh, A., Deo, R.C., Ali, M., Abdulla, S. and Raj, N. (2021) Novel Short-Term Solar Radiation Hybrid Model: Long Short-Term Memory Network Integrated with Robust Local Mean Decomposition. Applied Energy, 298, Article ID: 117193. https://doi.org/10.1016/j.apenergy.2021.117193 |
| [7] | 杨青, 王晨蔚. 基于深度学习LSTM神经网络的全球股票指数预测研究[J]. 统计研究, 2019, 36(3): 65-77. |
| [8] | 赵鑫, 陈臣鹏, 毕贵红, 等. 基于多模式分解的碳交易价格组合预测模型[J]. 电力科学与工程, 2022, 38(2): 52-59. |
| [9] | Yang, H., Yang, X. and Li, G. (2023) Forecasting Carbon Price in China Using a Novel Hybrid Model Based on Secondary Decomposition, Multi-Complexity and Error Correction. Journal of Cleaner Production, 401, Article ID: 136701. https://doi.org/10.1016/j.jclepro.2023.136701 |
| [10] | 郑雪文. 基于VMD-SE-PSO-LSTM的碳交易价格预测研究[D]: [硕士学位论文]. 郑州: 河南大学, 2023. |
