Financial Time Series Forecasting Method Based on Multi-Scale Features and Attention Mechanism

doi:10.3778/j.issn.1002-8331.2203-0138

Abstract

Abstract: Financial time series forecasting is a very important practical problem in the economic field. However, due to the noise and volatility of financial markets, the forecasting accuracy of current existing methods is not yet satisfactory. In order to improve the prediction accuracy of financial time series, a hybrid prediction model DCNN_LSTM_AT integrating dilated convolutional neural network（DCNN）, long short term memory（LSTM） and attention mechanism（AT） is proposed. The model consists of two parts：the first part contains a dilated convolutional neural network and an LSTM-based encoder, whose function is to extract effective information at different temporal scales in the original sequence data. The second part is composed of an LSTM decoder with an attention mechanism, its function is to filter the information extracted from the first part and use the filtered information to make predictions. Finally, the proposed model is tested on 3 index datasets and 3 individual stock datasets, and compared with other common benchmark models. The experimental results show that the model has better prediction accuracy and stability than other models.

Key words: stock index forecast, dilated convolutional neural network（DCNN）, attention mechanism（AT）, long short term memory（LSTM） neural network

摘要： 金融时间序列预测是经济领域中一个非常重要的实际问题，然而，由于金融市场的噪声和波动性，当前存在方法的预测精度尚不能令人满意。为了提高金融时间序列的预测精度，提出了一种融合扩张卷积神经网络（dilated convolutional neural network，DCNN）、长短时记忆神经网络（long short term memory，LSTM）和注意力机制（attention mechanism，AT）的混合预测模型DCNN_LSTM_AT。该模型由两个部分组成：第一部分包含扩张卷积神经网络和基于LSTM的编码器，其功能在于提取原始序列数据中不同时间尺度的有效信息；第二部分由带注意力机制的LSTM解码器构成，其功能在于对第一部分提取的信息进行过滤并利用过滤后的信息进行预测。最后将所提模型在3支股指数据集和3支个股数据集上进行实验，并与其他常见的基准模型进行了对比，实验结果表明该模型相比于其他模型具有更好的预测精度和稳定性。

关键词: 股指预测, 扩张卷积神经网络（DCNN）, 注意力机制, 长短时记忆神经网络

ZHAN Xi, PAN Zhisong, LI Wei, ZHANG Yanyan, BAI Wei WANG Cailing. Financial Time Series Forecasting Method Based on Multi-Scale Features and Attention Mechanism[J]. Computer Engineering and Applications, 2022, 58(19): 107-115.

詹熙, 潘志松, 黎维, 张艳艳, 白玮, 王彩玲. 基于多尺度特征和注意力的金融时序预测方法[J]. 计算机工程与应用, 2022, 58(19): 107-115.

References

[1] WANG B H，HUANG H J，WANG X L.A novel text mining approach to financial time series forecasting[J].Neurocomputing，2012，83（6）：136-145.
[2] SMOLA A J，SCH?LKOPF B.A tutorial on support vector regression[J].Statistics and Computing，2004，14（3）：199-222.
[3] TORRES J F，GALICIA A，TRONCOSO A，et al.A scalable approach based on deep learning for big data time series forecasting[J].Integrated Computer-Aided Engineering，2018，25（4）：335-348.
[4] ALEX K，ILYA S，GEOFFREY E H.ImageNet classification with deep convolutional neural networks[J].Communications of the ACM，2017，60（6）：84-90.
[5] CHRISTIAN S，LIU W，JIA Y Q，et al.Going deeper with convolutions[C]//Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition（CVPR），2015：1-9.
[6] DAHL G E.YU D，DENG L，et al.Context-dependent pre-trained deep neural networks for large-vocabulary speech recognition[J].IEEE Transactions on Audio，Speech，and Language，2012，20（1）：30-42.
[7] SUTSKEVER I，VINYALS O，LE Q V.Sequence to sequence learning with neural networks[C]//Advances in Neural Information Processing Systems，2014：3104-3112.
[8] BENGIO Y，DUCHARME R，VINCENT P，et al.A neural probabilistic language model[J].The Journal of Machine Learning Research，2003，3：1137-1155.
[9] MIKOLOV T，SUTSKEVER I，CHEN K，et al.Distributed representations of words and phrases and their compositionality[C]//Advances in Neural Information Processing Systems，2013：3111-3119.
[10] LI J H，QUN D，RUI Y.A novel double incremental learning algorithm for time series prediction[J].Neural Computing and Applications，2019，31（10）：6055-6077.
[11] ZHENG J，XIAO F，ZHANG G J.Research on exchange rate forecasting based on deep belief network[J].Neural Computing and Applications，2019，31（1）：573-582.
[12] SCHMIDHUBER J.Deep learning in neural networks：An overview[J].Neural Networks，2015，61：85-117.
[13] YANG Y，XIONG Q，WU C，et al.A study on water quality prediction by a hybrid CNN-LSTM model with attention mechanism[J].Environmental Science and Pollution Research，2021，28：55129-55139.
[14] BOX G E P，JENKINS G M，REINSEL G C，et al.Time series analysis：Forecasting and control[M].[S.l.]：John Wiley & Sons，2015.
[15] PETER ?，SILVIA P.ARIMA vs.ARIMAX—which approach is better to analyze and forecast macroeconomic time series[C]//Proceedings of the 30th International Conference on Mathematical Methods in Economics，2012：136-140.
[16] HORNIK K，STINCHCOMBE M，WHITE H.Multilayer feedforward networks are universal approximators[J].Neural Networks，1989，2（5）：359-366.
[17] HAMZA?EBI C，AKAY D，KUTAY F.Comparison of direct and iterative artificial neural network forecast approaches in multi-periodic time series forecasting[J].Expert Systems with Applications，2009，36（2）：3839-3844.
[18] NADUVIL-VADUKOOTU S，ANGRYK R A，RILEY P.Evaluating preprocessing strategies for time series prediction using deep learning architectures[C]//Proceedings of the Thirtieth International Flairs Conference，2017.
[19] CONNOR J T，MARTIN R D，ATLAS L E.Recurrent neural networks and robust time series prediction[J].IEEE Transactions on Neural Networks，1994，5（2）：240-254.
[20] HOCHREITER S，SCHMIDHUBER J.Long short-term memory[J].Neural Computation，1997，9（8）：1735-1780.
[21] XUE Y，WANG W H，MIAO C.Research on financial assets transaction prediction model based on LSTM neural network[J].Neural Computing and Applications，2021，33（1）：257-270.
[22] SHEN Z P，ZHANG Y M，LU J W，et al.A novel time series forecasting model with deep learning[J].Neurocomputing，2020，396：302-313.
[23] LU W，LI J，LI Y，et al.A CNN-LSTM-based model to forecast stock prices[J].Complexity，2020（1）：1-10.
[24] YU F，KOLTUN V.Multi-scale context aggregation by dilated convolutions[J].arXiv：1511.07122，2015.
[25] LARA-BENíTEZ P，CARRANZA-GARCíA M，RIQUELME J C.An experimental review on deep learning architectures for time series forecasting[J].arXiv：2103.12057，2021.
[26] BENGIO Y，SIMARD P，FRASCONI P.Learning long-term dependencies with gradient descent is difficult[J].IEEE Transactions on Neural Networks，1994，5（2）：157-166.
[27] BOUKTIF S，FIAZ A，OUNI A，et al.Optimal deep learning lstm model for electric load forecasting using feature selection and genetic algorithm：Comparison with machine learning approaches[J].Energies，2018，11（7）：1636.
[28] WU N，GREEN B，BEN X，et al.Deep transformer models for time series forecasting：The influenza prevalence case[J].arXiv：2001.08317，2020.
[29] LAI G，CHANG W C，YANG Y，et al.Modeling long-and short-term temporal patterns with deep neural networks[C]//Proceedings of the 41st International ACM SIGIR Conference on Research & Development in Information Retrieval，2018：95-104.