计算机科学

计算机科学 ›› 2025, Vol. 52 ›› Issue (10): 22-32.doi: 10.11896/jsjkx.250300104

• 数智赋能金融科技前沿 • 上一篇    下一篇

群组交叉对抗模型在股价预测中的应用

李奥1, 白雪茹2, 姜佳丽3, 乔烨4   

  1. 1 同济大学计算机科学与技术学院 上海 201804
    2 西北农林科技大学经济管理学院 陕西 咸阳 712100
    3 上海理工大学健康科学与工程学院 上海 200093
    4 同济大学电子与信息工程学院 上海 201804
  • 收稿日期:2025-03-20 修回日期:2025-06-13 出版日期:2025-10-15 发布日期:2025-10-14
  • 通讯作者: 乔烨(9@qiaoye.com)
  • 作者简介:(18554428089@outlook.com)

Group Cross Adversarial Application in Stock Price Prediction

LI Ao1, BAI Xueru2, JIANG Jiali3, QIAO Ye4   

  1. 1 School of Computer Science and Technology,Tongji University,Shanghai 201804,China
    2 School of Economics and Management,Northwest A&F University,Xianyang,Shaanxi 712100,China
    3 School of Health Science and Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China
    4 School of Electronics and Information Engineering,Tongji University,Shanghai 201804,China
  • Received:2025-03-20 Revised:2025-06-13 Online:2025-10-15 Published:2025-10-14
  • About author:LI Ao,born in 2003,undergraduate,is a member of CCF(No.Z2003G).His main research interests include machine learning and large meteorological models.
    QIAO Ye,born in 1973,Ph.D,is a member of CCF(No.Z2018G).His main research interests include artificial intelligence and quantitative finance.

PDF (PC)

5

摘要： 股票价格预测一直是金融研究和量化投资共同关注的重点话题。针对传统GAN模型存在模式崩溃与泛化能力弱的问题,同时为提高股价预测准确度,提出了群组交叉对抗模型(GCA),该模型包含多个生成器和多个判别器,同时在生成器和判别器间引入协作机制,以提升生成器的泛化能力,并通过知识蒸馏进一步提升生成器的预测性能。实验选取2015年1月1日至2025年1月1日期间A股(工商银行、华能国际、招商银行和青岛海尔)和美股(阿里巴巴、亚马逊、京东和美国银行)共8只股票的日度数据作为研究样本,构建了包括市场数据、技术指标在内的24个特征变量的数据集。研究结果表明,GCA模型在MAE,MAPE和MSE这3项评估指标上的表现明显优于单独应用的GRU,LSTM和Transformer模型,同时还优于结合了GAN的GRU-GAN,LSTM-GAN和Transformer-GAN模型,以及WGAN-GP和ResNLS模型;即使GAN并未对原始模型进行优化,但其引入GCA框架依旧提高了模型预测精度。进一步的讨论显示,增加生成器和判别器组数可以进一步提升预测效果。

关键词: 股价预测, 生成对抗网络, 时序预测模型, 多生成器多判别器模型

Abstract: Stock price prediction has always been a key topic of focus in both financial research and quantitative investment.To address the issues of mode collapse and weak generalization ability in traditional GAN models,and to improve stock price prediction accuracy,this study proposes a Group Cross-Adversarial model (GCA).The model includes multiple generators and multiple discriminators,and introduces a collaboration mechanism between the generators and discriminators to enhance the generalization ability of the generators.Additionally,knowledge distillation is used to further improve the prediction performance of the generators.The experiment selects daily data of 8 stocks (Industrial and Commercial Bank of China,Huaneng International,China Merchants Bank,and Qingdao Haier from the A-share market,and Alibaba,Amazon,JD,and Bank of America from the US stock market) from January 1,2015 to January 1,2025 as the research sample,and constructs a dataset of 24 feature variables,including market data and technical indicators.The results show that the proposed GCA model significantly outperforms the standalone GRU,LSTM,and Transformer models in terms of the three evaluation metrics－MAE,MAPE,and MSE.Additionally,it surpasses the GRU-GAN,LSTM-GAN,and Transformer-GAN models,which integrate GAN architectures,as well as the WGAN-GP and ResNLS models.Even without optimizing the original models with GAN,the introduction of the GCA framework still improves the prediction accuracy.Further discussion indicates that increasing the number of generator-discriminator pairs can further optimize prediction performance.

Key words: Stock price prediction,GAN,Time series prediction model,Multi-generator multi-discriminator model

中图分类号: 

  • TP183
[1]HAN Z M,MENG Y X,GUO H Y,et al.Stock price prediction based on dynamic heterogeneous networks[J].Computer Applications Research,2024,41(7):2126-2133.
[2]ZHANG Y,LI L.RF-MIP-GRU stock price prediction model[J].Computer Engineering and Applications,2024,60(17):272-281.
[3]LONG W,LU Z,CUI L.Deep learning-based feature enginee-ring for stock price movement prediction[J].Knowledge-Based Systems,2019,164:163-173.
[4]GOODFELLOW I,POUGET-ABADIE J,MIRZA M,et al.Ge-nerative adversarial nets[C]//Proceedings of the 27th International Conference on Neural Information Processing Systems.Cambridge,2014:2672-2680.
[5]WU H X,WANG Q,LI J P,et al.Enhancing stock return prediction in the Chinese market:A GAN-based approach[J].Research in International Business and Finance,2025,75:102760.
[6]SUN Z T,HARIT A,CRISTEA A I,et al.MONEY:Ensemble learning for stock price movement prediction via a convolutional network with adversarial hypergraph model[J].AI Open,2023,4:165-174.
[7]POLAMURI S R,SRINIVAS DR K,MOHAN DR A.Multi-Model Generative Adversarial Network Hybrid Prediction Algorithm(MMGAN-HPA) for stock market prices prediction[J].Journal of King Saud University-Computer and Information Sciences,2022,34(9):7433-7444.
[8]YANG M,WANG J.Research on stock index prediction basedon spatiotemporal attention mechanism and bidirectional long short-term memory neural network[J].Operations Research and Management,2023,32(8):174-180.
[9]ZHOU J N,LIU C Y,LIU J P.Stock price trend predictionmodel based on channel and multi-head attention[J].Computer Engineering and Applications,2025,61(8):324-338.
[10]NEJAD F S,EBADZADEH M M.Stock market forecastingusing DRAGAN and feature matching[J].Expert Systems with Applications,2023,244:122952.
[11]WANG Z X,WANG B,LI Y,et al.Cross-modal scenario gene-ration for stock price forecasting using Wasserstein GAN and GCN[J].Applied Soft Computing,2024,167(B):112342.
[12]SHEN R C,ZHAI J H,HOU Y Z.Multi-generator generative adversarial networks[J].Journal of Hebei University(Natural Science Edition),2021,41(6):734-744.
[13]LI W,LIANG Z,NEUMAN J,et al.Multi-generator GANlearning disconnected manifolds with mutual information[J].Knowledge-Based Systems,2021,212:106513.
[14]LI X H,WANG J,JIA H D,et al.Stock market volatility prediction method based on multi-attention mechanism and graph neural networks[J].Computer Applications,2022,42(7):2265-2273.
[15]WANG K F,GOU C,DUAN Y J,et al.Research progress and prospects of generative adversarial network(GAN)[J].Acta Automatica Sinica,2017,43(3):321-332.
[16]ZHENG X B,LI J,ZHU M,et al.Cloud manufacturing industrial service selection method based on generative adversarial network[J].Computer Science,2025,52(4):54-63.
[17]ZOU J,LI L.Research on stock price prediction based on RF-SA-GRU model[J].Computer Engineering and Applications,2023,59(15):300-309.
[18]HU Y W.Stock prediction based on optimized LSTM model[J].Computer Science,2021,48(S1):151-157.
[19]CHENG H Y,ZHANG J X,SUN Q S,et al.Stock trend prediction based on deep cross-modal information fusion network[J].Computer Science,2023,50(5):128-136.
[20]FANG Y Q,LU Z,GE J W.Stock price prediction using joint RMSE loss LSTM-CNN model[J].Computer Engineering and Applications,2022,58(9):294-302.
[21]CAO C F,LUO Z N,XIE J X,et al.Research on stock price prediction based on MDT-CNN-LSTM model[J].Computer Engineering and Applications,2022,58(5):280-286.
[22]SHAO R R,LIU Y A,ZHANG W,et al.A survey on knowledge distillation in deep learning[J].Journal of Computer Science,2022,45(8):1638-1673.
[23]HUANG Z H,YANG S Z,LIN W,et al.A survey on knowledge distillation[J].Journal of Computer Science,2022,45(3):624-653.
[24]LIN H C,CHEN C,HUANG G,et al.Stock price predictionusing Generative Adversarial Networks[J].Journal of Computer Science,2021,17(3):188-196.
[25]JIA Y,ANAISSI A,SULEIMAN B.ResNLS:An ImprovedModel for Stock Price Forecasting[J].Computational Intelligence,2023,40(1):12608.
[26]CHANG W,HU Z C,PAN D Z,et al.Application of improved VAE-GAN model in battery EIS data augmentation[J].Science and Industry,2024,24(22):258-263.
[27]GE Y B,LIU W J,GU Y C.Stock price prediction method integrating sentiment analysis and GAN-TrellisNet[J].Computer Engineering and Applications,2024,60(12):314-324.
[28]CHANG D L,YIN J H,XIE J Y,et al.Attention-guided Dropout for image classification[J].Journal of Graphics,2021,42(1):32-36.
[29]WANG L,YANG J,ZHANG C Y,et al.Dual-discriminator generative adversarial network with hybrid attention[J].Computer Engineering and Applications,2024,60(7):212-221.
[30]XIA Y S,WANG Y,ZHOU L,et al.Fake data injection attack detection method based on improved generative adversarial network[J].Electric Power Construction,2022,43(3):58-65.
[31]WANG P Y,ZHU Z Q.Decoupled knowledge distillation based on diffusion model[J].Computer Systems Applications,2024,33(9):58-64.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发