《计算机应用》唯一官方网站 ›› 2022, Vol. 42 ›› Issue (3): 797-803.DOI: 10.11772/j.issn.1001-9081.2021050748

• 2021年中国计算机学会人工智能会议(CCFAI 2021) • 上一篇    

基于时序超图卷积神经网络的股票趋势预测方法

李晓杰1, 崔超然1(), 宋广乐2, 苏雅茜1, 吴天泽3, 张春云1   

  1. 1.山东财经大学 计算机科学与技术学院, 济南 250014
    2.山东省人工智能学会, 济南 250101
    3.齐鲁工业大学(山东省科学院) 计算机科学与技术学院, 济南 250353
  • 收稿日期:2021-05-11 修回日期:2021-07-16 接受日期:2021-07-21 发布日期:2021-11-09 出版日期:2022-03-10
  • 通讯作者: 崔超然
  • 作者简介:李晓杰(1998—),女,山东聊城人,硕士研究生,主要研究方向:深度学习、数据挖掘
    宋广乐(1992—),男,山东德州人,硕士,主要研究方向:计算机视觉、机器学习、人工智能
    苏雅茜(1997—),女,山东济南人,硕士研究生,主要研究方向:深度学习、数据挖掘
    吴天泽(1991—),男,山东济南人,讲师,硕士,主要研究方向:数据挖掘
    张春云(1986—),女,山东聊城人,副教授,博士,主要研究方向:自然语言处理、机器学习。
  • 基金资助:
    国家自然科学基金资助项目(62077033);国家重点研发计划项目(2018YFC0830100);山东省自然科学基金重点项目(ZR2020KF015);山东省高等学校优势学科人才团队培育计划项目

Stock trend prediction method based on temporal hypergraph convolutional neural network

Xiaojie LI1, Chaoran CUI1(), Guangle SONG2, Yaxi SU1, Tianze WU3, Chunyun ZHANG1   

  1. 1.School of Computer Science and Technology,Shandong University of Finance and Economics,Jinan Shandong 250014,China
    2.Shandong Association for Artificial Intelligence,Jinan Shandong 250101,China
    3.School of Computer Science and Technology,Qilu University of Technology (Shandong Academy of Sciences),Jinan Shandong 250353,China
  • Received:2021-05-11 Revised:2021-07-16 Accepted:2021-07-21 Online:2021-11-09 Published:2022-03-10
  • Contact: Chaoran CUI
  • About author:LI Xiaojie, born in 1998, M. S. candidate. Her research interests include deep learning, data mining.
    SONG Guangle, born in 1992, M. S. His research interests include computer vision, machine learning, artificial intelligence.
    SU Yaxi, born in 1997, M. S. candidate. Her research interests include deep learning, data mining.
    WU Tianze, born in 1991, M. S., lecturer. His research interests include data mining.
    ZHANG Chunyun, born in 1986, Ph. D., associate professor. Her research interests include natural language processing, machine learning.
  • Supported by:
    National Natural Science Foundation of China(62077033);National Key Research and Development Program of China(2018YFC0830100);Shandong Provincial Natural Science Foundation Key Project(ZR2020KF015);Fostering Project of Dominant Discipline and Talent Team of Shandong Province Higher Education Institutions

摘要:

传统的股票预测方法大多基于时间序列模型,忽视了股票之间复杂的关系,并且该关系往往超出成对连接,例如同行业板块内股票或者基金持仓多支股票。针对该问题,提出一种基于时序超图卷积神经网络(HGCN)的股价走势预测方法,根据金融投资事实构造超图模型以拟合股票之间的多元关系,该模型包括两大组件:门控循环单元(GRU)网络和超图卷积神经网络。GRU网络对历史数据进行时间序列建模,捕捉长期依赖关系;HGCN建模股票间的高阶关系以学习内在关系属性,从而将股票间多元关系信息引入到传统的时序建模中,进行端到端的趋势预测。在中国A股市场真实数据集上的实验结果表明,相较于已有的股票预测方法,所提模型预测性能有所提升;如与GRU网络相比,所提模型在ACC和F1_score上的相对增幅分别为9.74%和8.13%,且更具有稳定性。此外,模拟回测结果显示,基于该模型的交易策略更具获利能力,年回报率达到11.30%,与长短期记忆(LSTM)网络相比提高了5个百分点。

关键词: 股票趋势预测, 时间序列建模, 门控循环单元, 高阶关系, 超图卷积神经网络

Abstract:

Traditional stock prediction methods are mostly based on time-series models, which ignore the complex relations among stocks, and the relations often exceed pairwise connections, such as stocks in the same industry or multiple stocks held by the same fund. To solve this problem, a stock trend prediction method based on temporal HyperGraph Convolutional neural Network (HGCN) was proposed, and a hypergraph model based on financial investment facts was constructed to fit multiple relations among stocks. The model was composed of two major components: Gated Recurrent Unit (GRU) network and HGCN. GRU network was used for performing time-series modeling on historical data to capture long-term dependencies. HGCN was used to model high-order relations among stocks to learn intrinsic relation attributes, and introduce the multiple relation information among stocks into traditional time-series modeling for end-to-end trend prediction. Experiments on real dataset of China A-share market show that compared with existing stock prediction methods, the proposed model improves prediction performance, e.g. compared with the GRU network, the proposed model achieves the relative increases in ACC and F1_score of 9.74% and 8.13%, respectively, and is more stable. In addition, the simulation back-testing results show that the trading strategy based on the proposed model is more profitable, with an annual return of 11.30%, which is 5 percentage points higher than that of Long Short-Term Memory (LSTM) network.

Key words: stock trend prediction, time series modeling, Gated Recurrent Unit (GRU), high-order relation, HyperGraph Convolutional neural Network (HGCN)

中图分类号: