基于注意力时间卷积网络的农产品期货分解集成预测
作者单位:

华南农业大学

基金项目:

国家自然科学基金面上项目(71971089);国家自然科学基金青年项目(72001083);广东省自然科学基金面上项目(2022A1515011612)


Forecasting the agricultural futures with decomposition-ensemble strategy based on attentional temporal convolution network
Author:
Affiliation:

South China Agricultural University

  • 摘要
  • | |
  • 访问统计
  • |
  • 参考文献
  • | |
  • 引证文献
  • | |
  • 文章评论
    摘要:

    针对农产品期货时间序列数据受多方面因素影响,非线性、非平稳数据特征难以提取而导致预测准确性不高的问题,本文基于“分解-集成”的预测思想,提出一种基于CEEMDAN与TransformerEncoder-TCN的农产品期货预测方法。首先,使用自适应噪声完备经验模态分解(CEEMDAN)将时间序列分解为多尺度多频率的本征模态分量(Intrinsic Mode Function, IMF)与残差,降低了序列建模复杂度;其次,使用融合多阶段自注意力单元TransformerEncoder的时间卷积网络(Temporal Convolution Network, TCN)对各个分量子序列进行特征提取与预测,优化了序列显著特征建模权重;最后,将各个子序列预测值线性相加集成得到最终预测结果。以南华期货公司农产品指数中的大豆期货指数为研究对象,采用时序交叉验证与参数迁移的方式进行模型重训练,消融和对比实验结果表明,提出的新模型在RMSE、MAE和DS三个评价指标上具有良好的效果,验证了该模型对农产品期货预测的有效性。

    Abstract:

    In response to the problem that agricultural futures are affected by many factors so that nonlinear and nonsmooth features are difficult to extract and lead to low prediction accuracy, this paper proposes a forecasting method based on CEEMDAN and TransformerEncoder-TCN, which is from the idea of "decomposition-ensemble". First, the complementary ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is used to decompose the time series into multiscale Intrinsic Mode Function (IMF) and residuals, reducing the complexity of series modeling. Second, each subseries is predicted using Temporal Convolutional Network (TCN) incorporating multi-stage self-attention unit (TransformerEncoder) which optimizes the modeling weights of the significant features of the sequence. Finally, the prediction results of each subseries are linearly summed and integrated to obtain the final prediction results. The soybean futures revenue index in the agricultural product index of South China Futures Company is used as the research object. The model is retrained by using time-series cross-validation and parameter transfer. The ablation and comparison experimental results show that the model proposed in this paper has superiority in RMSE, MAE and DS, verifying its effectiveness in predicting agricultural product futures.

    参考文献
    [1] 刘洋,罗其友,周振亚,等. 我国主要农产品供需分析与预测[J], 中国工程科学. 2018, 20(05): 120-127.IU Yang, LUO Qiyou, ZHOU Zhengya, et al. Analysis and Prediction of the Supply and Demand of China’s Major Agricultural Products[J]. Strategic Study of CAE, 2018, 20(05):120-127.
    [2] Z. Han, J. Zhao, H. Leung,et al. A Review of Deep Learning Models for Time Series Prediction[J]. IEEE Sensors Journal, 2019, 21(6): 7833-7848.
    [3] 张永安,颜斌斌. 一种股票市场的深度学习复合预测模型[J], 计算机科学. 2020, 47(11): 255-267.HANG Yongan, YAN Binbin. Deep Learning Hybrid Forecasting Model for Stock Markets[J]. Computer Science, 2020, 47(11):255-267.
    [4] 赵娜,孙红,黎铨祺,等. 基于小波分解的时序预测模型及其应用研究[J]. 小型微型计算机系统, 2021: 1-9.HAO Na, SUN Hong, LI Quanqi, et al. Research on Application of Time Series Forecast Model Based on Wavelet? Decomposition[J]. Journal of Chinese Computer Systems, 2021,1-9.
    [5] Wang S, Mu L, Liu D. A hybrid approach for El Ni?o prediction based on Empirical Mode Decomposition and convolutional LSTM Encoder-Decoder[J]. Computers Geosciences, 2021, 149(2):104695.
    [6] Wang Y, Yuan Z, Liu H, et al. A new scheme for probabilistic forecasting with an ensemble model based on CEEMDAN and AM-MCMC and its application in precipitation forecasting[J]. Expert Systems with Applications, 2022, 187(4):115872.
    [7] 唐振鹏,吴俊传,张婷婷,等. 基于二次分解和集成学习的粮食期货价格预测研究[J], 系统工程理论与实践. 2021: 1-16.ANG Zhengpeng, WU Junchuan, ZHANG Tingting, et al. Research on Grain Futures Price Forecasting Based on Secondary Decomposition and Ensemble Learning[J]. Systems Engineering-Theory Practice, 2021,1-16.
    [8] Zhou Y, Li T, Shi J, et al. A CEEMDAN and XGBOOST-Based Approach to Forecast Crude Oil Prices[J]. Complexity, 2019,(2019): 1-15.
    [9] 韩金磊,熊萍萍,孙继红.基于LSTM和灰色模型的股价时间序列预测研究[J/OL].南京信息工程大学学报(自然科学版):1-22[2023-01-06].http://kns.cnki.net/kcms/detail/32.1801.N.20230105.1635.003.html.HAN Jinlei, XIONG Pingping,SUN Jihong.Stock price time series prediction based on LSTM and greymodel[J].Journal of Nanjing University of Information Science Technology(Natural ScienceEdition) :1-22[2023-01-06].http://kns.cnki.net/kcms/detail/32.1801.N.20230105.1635.003.html
    [10] BAI Shaojie, Kolter Zico J,Koltun Vladlen. An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling[J]. arXiv: 1803.0127,2018
    [11] Li G, Chen W, Li D, et al. Comparative Study of Short-Term Forecasting Methods for Soybean Oil Futures Based on LSTM, SVR, ES and Wavelet Transformation[J]. Journal of physics. Conference series, 2020, 1682(1).
    [12] Cheng W, Wang Y, Peng Z, et al. High-efficiency chaotic time series prediction based on time convolution neural network[J]. Chaos, Solitons Fractals, 2021, (152):111304.
    [13] Wang C, Gao Q. High and Low Prices Prediction of Soybean Futures with LSTM Neural Network[C]. China IEEE Computer Society, Beijing, 2018,140-143.
    [14] Chuluunsaikhan T, Ryu G, Yoo K, et al. Incorporating Deep Learning and News Topic Modeling for Forecasting Pork Prices The Case of South Korea[J]. Agriculture, 2020, 10(11).
    [15] 张杰,甄柳琳,徐硕,等. 融合传递熵的图神经网络农产品期货预测模型[J], 计算机工程与应用. 2021: 1-10.HANG Jie, ZHEN Liulin, XU Shuo, et al. Graph Neural Network Model Based on Transfer Entropy for Agricultural Futures Forecasting[J].Computer Engineering and Applications, 2021,1-10.
    [16] Wei X, Lei B, Ouyang H, et al. Stock Index Prices Prediction via Temporal Pattern Attention and Long-Short-Term Memory[J]. Advances in Multimedia, 2020, (2020): 1-7
    [17] Mnih V, Heess N, Graves A, et al. Recurrent Models of Visual Attention[C]. Advances in Neural Information Processing Systems, 2014, 2204-2212.
    [18] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[C]. Neural Information Processing Systems Foundation, Long Beach, CA, 2017,5999-6009.
    [19] Niu H, Xu K, Liu C. A decomposition-ensemble model with regrouping method and attention-based gated recurrent unit network for energy price prediction[J]. Energy, 2021, (231): 120941.
    [20] 贾宁,郑纯军. 基于LSTM-DA神经网络的农产品价格指数短期预测模型[J], 计算机科学. 2019, 46(S2): 62-65.IA Ning, ZHENG Chunjun. Short-term Forecasting Model of Agricultural Product Price Index Based on LSTM-DA Neural Network[J].Computer Science, 2019, 46(S2):62-65.
    [21] Yin H, Jin D, Gu Y H, et al. STL-ATTLSTM Vegetable Price Forecasting Using STL and Attention Mechanism-Based LSTM[J]. Agriculture, 2020, 10(12).
    [22] M. E.Torres, M. A. Colominas, G. Schlotthaue, et al. A complete ensemble empirical mode decomposition with adaptive noise[C]. 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2011, pp.4144-4147.
    [23] Islam M A, Jia S, Bruce N. How Much Position Information Do Convolutional Neural Networks Encode[J]. ArXiv:2001.08248,2020.
    [24] Wang X , Girshick R , Gupta A , et al. Non-local Neural Networks[C]. IEEECVF Conference on Computer Vision and Pattern Recognition, 2018, 7794-7803.
    [25] Nair B B, Mohandas V P, Sakthivel N R. A Decision Tree-Rough Set Hybrid System for Stock Market Trend Prediction[J]. International Journal of Computer Applications, 2010, 6(9):1-6.
    [26] Chan Phooi M''Ng J, Mehralizadeh M, Gao Z. Forecasting East Asian Indices Futures via a Novel Hybrid of Wavelet-PCA Denoising and Artificial Neural Network Models[J]. PloS one, 2016, 11(6): e0156338.
    相似文献
    引证文献
    网友评论
    网友评论
    分享到微博
    发 布
引用本文

张大斌,黄均杰,凌立文,林锐斌.基于注意力时间卷积网络的农产品期货分解集成预测[J].南京信息工程大学学报,,():

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-08-22
  • 最后修改日期:2023-10-20
  • 录用日期:2023-10-27

地址:江苏省南京市宁六路219号    邮编:210044

联系电话:025-58731025    E-mail:nxdxb@nuist.edu.cn

南京信息工程大学学报 ® 2025 版权所有  技术支持:北京勤云科技发展有限公司