计算机科学 ›› 2023, Vol. 50 ›› Issue (5): 128-136.doi: 10.11896/jsjkx.220400089

• 数据库&大数据&数据科学 • 上一篇    下一篇

基于深度跨模态信息融合网络的股票走势预测

程海阳1, 张建新2, 孙启森1, 张强1,3, 魏小鹏1,3   

  1. 1 大连大学先进设计与智能计算教育部重点实验室 辽宁 大连 116622
    2 大连民族大学计算机科学与工程学院 辽宁 大连 116600
    3 大连理工大学计算机科学与技术学院 辽宁 大连 116024
  • 收稿日期:2022-04-11 修回日期:2022-09-13 出版日期:2023-05-15 发布日期:2023-05-06
  • 通讯作者: 张强(zhangq@dlut.edu.cn)
  • 作者简介:(chenghaiyang@s.dlu.edu.cn)
  • 基金资助:
    国家自然科学基金辽宁省联合基金(U1908214);国家自然科学基金(61972062);辽宁省“兴辽英才计划”项目(XLYC2008017);辽宁省重点研发计划(2019JH2/10100030)

Deep Cross-modal Information Fusion Network for Stock Trend Prediction

CHENG Haiyang1, ZHANG Jianxin2, SUN Qisen1, ZHANG Qiang1,3, WEI Xiaopeng1,3   

  1. 1 Ministry of Education Key Laboratory of Advanced Design and Intelligent Computing,Dalian University,Dalian,Liaoning 116622,China
    2 School of Computer Science and Engineering,Dalian Minzu University,Dalian,Liaoning 116600,China
    3 School of Computer Science and Technology,Dalian University of Technology,Dalian,Liaoning 116024,China
  • Received:2022-04-11 Revised:2022-09-13 Online:2023-05-15 Published:2023-05-06
  • About author:CHENG Haiyang,born in 1998,master,is a member of China Computer Federation.His main research interests include data mining,series forecasting,and machine learning.
    ZHANG Qiang,born in 1971,Ph.D,professor,Ph.D supervisor,is a member of China Computer Federation.His main research interests include Biocomputing and artificial intelligence,and intelligent big data processing.
  • Supported by:
    National Natural Science Foundation of Liaoning Provincial Joint Fund(U1908214),National Natural Science Foundation of China(61972062),Liaoning Revitalization Talents Program(XLYC2008017) and Liaoning Province Key R&D Program(2019JH2/10100030).

摘要: 股票走势预测是经典且具有挑战性的任务,可帮助交易者做出获得更大收益的交易决策。近年来,基于深度学习的股票走势预测方法的性能得到明显提升,但现有方法大多仅依托于股票价格的历史数据来完成走势预测,无法捕捉价格指标之外的市场动态规律,在一定程度上限制了方法的性能。为此,将社交媒体文本与股票历史价格信息相结合,提出了一种基于深度跨模态信息融合网络(DCIFNet)的股票走势预测新方法。DCIFNet首先采用时间卷积操作对股票价格和推特文本进行编码,使得每个元素对其邻域元素都有足够的了解;然后,将结果输入到基于transformer的跨模态融合结构中,以更有效地融合股票价格和推特文本中的重要信息;最后,引入多图卷积注意力网络从不同角度描述不同股票之间的相互关系,能够更有效地捕获关联股票间的行业、维基和相关关系,从而提升股票走势预测的精度。在9个不同行业的高频交易数据集上实施走势预测和模拟交易实验。消融实验及所提方法与用于股票预测的多管齐下的注意力网络(MAN-SF)方法的比较结果验证了DCIFNet方法的有效性,准确率达到了 0.6309,明显优于领域内代表性方法。

关键词: 股票走势预测, 社交媒体文本, 跨模态信息融合, 图卷积网络, 时间卷积

Abstract: Stock trend prediction,as a classic and challenging task,can help traders make trading decisions for greater returns.Recently,deep learning related models have achieved obvious performance improvement on this task.However,most of the current deep learning related works only leverage the historical data on stock price to complete the trend prediction,which cannot capture the market dynamics other than price indicators,thus having an accuracy limitation to a certain extent.To this end,this paper combines social media texts with stock historical price information,and proposes a novel deep cross-modal information fusion network(DCIFNet) for stock trend prediction.DCIFNet first utilizes temporal convolution operations to encode stock prices and twitter texts,so that each element can have sufficient knowledge of its neighborhood elements.Then,the results are fed into a transformer-based cross-modal fusion structure to fuse stock prices and important information in Twitter texts more effectively.Finally,a multi-graph attention convolutional network is introduced to describe the interrelationships among different stocks,which well captures the industry,wiki and correlation relationship among related stocks,leading to the accuracy improvement of stock prediction.We have performed trend prediction and simulated trading experiments on high-frequency trading datasets in nine different industries,and ablation studies as well as compared experiments with multipronged attention network for stock forecasting(MAN-SF) demonstrate the effectiveness of the proposed DCIFNet method.In addition,with the optimal accuracy of 0.6309,it obviously outperforms representative methods on the stock prediction application.

Key words: Stock trend prediction, Social media text, Cross-modal information fusion, Graph convolutional network, Temporal convolution

中图分类号: 

  • TP391
[1]DEVI B U,SUNDAR D,ALLI P.An effective time series analysis for stock trend prediction using ARIMA model for nifty midcap-50[J].International Journal of Data Mining Knowledge Management Process,2013,3(1):65-78.
[2]DENG S M,ZHANG N Y,ZHANG W,et al.Knowledge-driven stock trend prediction and explanation via temporal convolu-tional network[C]//Companion Proceedings of The 2019 World Wide Web Conference.San Francisco,New York,USA:ACM,2019:678-685.
[3]HU Z N,LIU W Q,BIAN J,et al.Listening to chaotic whispers:A deep learning framework for news-oriented stock trend prediction[C]//Proceedings of the 18th ACM International Conference on Web Search and Data Mining.Marina Del Rey,New York,USA:ACM,2018:261-269.
[4]WENG B,AHMED M A,MEGAHED F M.Stock market one-day ahead movement prediction using disparate data sources[J].Expert Systems with Applications,2017,79(2017):153-163.
[5]HUANG J Y,ZHANG Y J,ZHANG J L,et al.A tensor-based sub-mode coordinate algorithm for stock prediction[C]//IEEE 3rd International Conference on Data Science in Cyberspace(DSC).Guangzhou,China.Piscataway:IEEE,2018:716-721.
[6]SAWHNEY R,AGARWAL S,WADHWA A,et al.Deep attentive learning for stock movement prediction from social media text and company correlations[C]//Proceedings of the 2020 Conference on Empirical Methods in Natural Language Proces-sing(EMNLP).Online.New York:ACM,2020:8415-8426.
[7]TSAI Y H H,BAI S J,LIANG P P,et al.Multimodal transformer for unaligned multimodal language sequences[C]//Proceedings of the Conference Association for Computational Linguistics Meeting.Bethesda:NIH Public Access,2019:6558-6569.
[8]CHEN L,CHI Y G,GUAN Y Y,et al.A Hybrid Attention-Based EMD-LSTM Model for Financial Time Series Prediction[C]//2nd International Conference on Artificial Intelligence and Big Data(ICAIBD).Chengdu,China.Piscataway:IEEE,2019:113-118.
[9]LIU M,SHAN Y Y.Predicition of Closing Price of Stock Index Based on EMD-LSTM Model[J].Journal of Chongqing University of Technology(Natural Science),2021,35(12):269-276.
[10]YANG B,GONG Z,YANG W.Stock market index predictionusing deep neural network ensemble[C]//2017 36th Chinese Control Conference(CCC).Dalian,China.Piscataway:IEEE,2017:3882-3887.
[11]CHENG L C,HUANG Y H,WU M E.Applied attention-based LSTM neural networks in stock prediction[C]//IEEE International Conference on Big Data(Big Data).Seattle,WA,USA.Piscataway:IEEE,2018 :4716-4718.
[12]BOLLERSLEV T.Generalized autoregressive conditional he-teroskedasticity[J].Journal of Econometrics,1986,31(3):307-327.
[13]ETHEM A.Introduction to machine learning[M].Cambridge:MIT press,2020.
[14]SANBOON T,KEATRUANGKAMALA K,JAIYEN S.ADeep Learning Model for Predicting Buy and Sell Recommendations in Stock Exchange of Thailand using Long Short-Term Memory[C]//2019 IEEE 4th International Conference on Computer and Communication Systems(ICCCS).Singapore.Piscataway:IEEE,2019:757-760.
[15]WANG D,WANG X P,YANG C D.A Study of Stock Forecasting Based on LSTM Model of Principal Component Analysis[J].Journal of Chongqing University of Technology(Natural Science),2021,35(2):282-288.
[16]ALEXIEI D,KARL F.Financial time series forecasting-a deep learning approach[J].International Journal of Machine Learning Computing,2017,7(5):118-122.
[17]LIU Q K,CHENG X,SU S,et al.Hierarchical Complementary Attention Network for Predicting Stock Price Movements with News[C]//Proceedings of the 27th ACM International Confe-rence on Information and Knowledge Management.Torino,Italy; ACM.2018:1603-1606.
[18]DING X,ZHANG Y,LIU T,et al.Deep learning for event-dri-ven stock prediction[C]//Proceedings of the 24th International Conference on Artificial Intelligence.Buenos Aires,Argentina:AAAI.2015:2327-2333.
[19]NIKFARJAM A,EMADZADEH E,MUTHAIYAH S.Textmining approaches for stock market prediction[C]//Interna-tional Conference on Computer and Automation Engineering.Singapore.Piscataway:IEEE,2010:256-260.
[20]WU H Z,ZHANG W,SHEN W W,et al.Hybrid deep sequential modeling for social text-driven stock prediction[C]//Proceedings of the 27th ACM International Conference on Information and Knowledge Management.Torino Italy.New York:ACM,2018:1627-1630.
[21]FENG F L,HE X N,WANG X,et al.Temporal relational ran-king for stock prediction[J].ACM Transactions on Information Systems,2019,37(2):1-30.
[22]CHEN Y M,WEI Z Y,HUANG X J.Incorporating corporation relationship via graph convolutional neural networks for stock price prediction[C]//ACM International Conference on Information and Knowledge Management.New York,United States.New York:ACM,2018:1655-1658.
[23]MATSUNAGA D,SUZUMURA T,TAKAHASHI T.Exploring graph neural networks for stock market predictions with rolling window analysis[J].arXiv:1909.10660,2019.
[24]KIM R,SO C H,JEONG M,et al.Hats:A hierarchical graph attention network for stock movement prediction[J].arXiv:1908.07999,2019.
[25]XU W T,LIU W Q,XU C,et al.REST:Relational Event-driven Stock Trend Forecasting[C]//Proceedings of the Web Confe-rence.Ljubljana,Slovenia.New York:ACM,2021:1-10.
[26]SERGEY L,CHRISTIAN S.Batch normalization:Accelerating deep network training by reducing internal covariate shift[C]//International Conference on Machine Learning.Lille,France.New York:PMLR,2015:448-456.
[27]XU Y,COHEN S B.Stock movement prediction from tweets and historical prices[C]//Proceedings of the 56th Annual Mee-ting of the Association for Computational Linguistics(Volume 1:Long Papers)Melbourne,Australia.Melbourne:ACL,2018:1970-1979.
[28]LEI B J,RYAN K J,E H G.Layer normalization[J].arXiv:1607.06450,2016.
[29]LIU G,WANG X J,LI R F.Multi-scale RCNN model for financial time-series classification[J].arXiv:1911.09359,2019.
[30]CHEN Q K,ROBERT C Y.Graph-Based Learning for Stock Movement Prediction with Textual and Relational Data[J].ar-Xiv:2107.10941,2021.
[31]YE J X,ZHAO J J,YE K J,et al.Multi-graph convolutional network for relationship-driven stock movement prediction[C]//2020 25th International Conference on Pattern Recognition(ICPR).Milan,Italy.Piscataway:IEEE,2021:6702-6709.
[32]BENESTY J,CHEN J D,HUANG Y T,et al.Pearson correlation coefficient[M]//Noise Reduction in Speech Processing.Berlin,Heidelberg:Springer.2009:1-4.
[33]KIPF T N,WELLINGS M.Semi-supervised classification with graph convolutional networks[J].arXiv:1609.02907,2016.
[34]GLOROT X,BORDES A,BENGIO Y.Deep sparse rectifierneural networks[C]//Proceedings of the 14th International Conference on Artificial Intelligence and Statistics.Microtome Publishing:JMLR Workshop and Conference Proceedings,2011:315-323.
[35]SHARPE W F.The sharpe ratio[J].Journal of portfolio ma-nagement,1998,21(1):169-185.
[36]CHICCO D,JURMAN G.The advantages of the Matthews correlation coefficient(MCC) over F1 score and accuracy in binary classification evaluation[J].BMC Genomics,2020,21(1):1-13.
[37]NICOLAS H.Pairs selection and outranking:An application to the S&P 100 index[J].European Journal of Operational Research,2009,196(2):819-825.
[38]NICOLAS H.Pairs trading and outranking:The multi-step-ahead forecasting case[J].European Journal of Operational Research,2010,207(3):1702-1716.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!