Strategy Teaching Research Based on Multilayer Graph Relationship and Reinforcement Learning

doi:10.19678/j.issn.1000-3428.0069277

Abstract

Abstract:

Developments in educational informatization have provided additional schemes for promoting the diversification of teaching content. To enrich the connotation of the programmatic trading classroom, a new case- teaching method is proposed to discuss and study quantitative strategies. In stock trend forecasting, the prediction of a company's stock is affected by the multifaceted, invisible level of executive relationships within the associated company. To effectively deal with the momentum spillover effect in stock market volatility, this study proposes a multilayer Graph Convolutional Neural Network (GCNN)-based model for stock trend forecasting and intelligent quantitative trading, considering the influence of executive factors. The model is applied in practical classroom teaching to enrich classroom examples used in teaching tasks. First, historical stock data and market media information are combined. Subsequently, explicit relationships between companies and implicit relationships among executives are fused, using a multilayer GCNN. Finally, strategy training is conducted through Reinforcement Learning (RL). This model not only effectively improves the accuracy of stock trend predictions but also promotes portfolio optimization profits. In experiments conducted on the CSI100E and CSI300E datasets, the proposed model is compared with existing state-of-the-art models, obtaining accuracies of 60.19% and 57.44%, respectively. In contrast, the Graph Convolutional Network (GCN) obtained accuracies of 51.58% and 55.79%, respectively. The analysis concluded that the stock trend predictions of the proposed model are more accurate, and intelligent investment decisions based on these predictive values are also more effective. The methodology and experimental results of this study provide real-world examples for financial-technology courses, helping students understand complex market dynamics and quantitative trading strategies.

Key words: Graph Neural Network (GNN), Deep Reinforcement Learning (DRL), stock trend prediction, quantitative investment, teaching case

摘要：

教育信息化的发展为促进教学内容多样化提供了更多方案。为了丰富程序化交易课程的内涵, 对量化策略的探讨研究提出了新的案例教学方法。由于在股票趋势预测的任务中, 公司股票的预测会受到相关联公司及多方面的隐性层面高管关系的影响。为了有效应对股票市场波动中的动量溢出效应, 就高管因素的影响提出一个基于多层图卷积神经网络(GCNN)的股票趋势预测及智能量化交易模型, 将其应用于实际课堂教学中充实课堂实例教学任务。该模型首先结合股票历史数据和市场媒体信息, 然后利用多层GCNN提取股票之间包含的具有交叉效应的公司间显性关系和高管间隐性关系等信息进行趋势预测, 最后通过强化学习(RL)进行策略训练。该模型不仅有效提高了股票趋势预测的准确性, 而且有效提升了投资组合优化收益。在CSI100E和CSI300E数据集上的实验结果表明, 该模型得到60.19%和57.44%的准确率, 而图卷积网络(GCN)模型得到51.58%和55.79%的准确率。通过分析得出该模型的股票趋势预测效果更好, 加入了预测结果的智能投资决策也更有效。该研究的方法和实验结果为金融课程提供了实际案例, 可帮助学生理解复杂市场动态和量化交易策略。

关键词: 图神经网络, 深度强化学习, 股票趋势预测, 量化投资, 教学案例

LI Siyuan, ZHONG Xingyu, LI Kaiyin, XU Qingzhen. Strategy Teaching Research Based on Multilayer Graph Relationship and Reinforcement Learning[J]. Computer Engineering, 2025, 51(3): 122-130.

李思源, 钟兴宇, 李凯茵, 徐清振. 基于多层图关系和强化学习的策略教学研究[J]. 计算机工程, 2025, 51(3): 122-130.

/ Recommend / Download Citations

URL: https://www.ecice06.com/EN/10.19678/j.issn.1000-3428.0069277

https://www.ecice06.com/EN/Y2025/V51/I3/122

Figures/Tables 8

Fig.1 Schematic diagram of general framework

Fig.2 DRL process framework

Fig.3 Schematic diagram of segmentation of experimental dataset

Fig.4 Profit comparison chart of single stock 000001 in the CSI300E dataset

Fig.5 Profit comparison chart for multiple stock portfolios in the CSI300E dataset

Fig.6 Profit comparison chart of single stock 600900 in the CSI100E dataset

Fig.7 Profit comparison chart for multiple stock portfolios in the CSI100E dataset

References 28

1	熊海芳. 量化投资分析在证券投资教学中的应用. 金融教学与研究, 2014 (2): 63- 66.
	XIONG H F . The application of quantitative investment analysis in securities investment teaching. Exploration of Financial Theory, 2014 (2): 63- 66.
2	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[EB/OL]. [2023-12-07]. http://arxiv.org/abs/1706.03762.
3	FENG F L , HE X N , WANG X , et al. Temporal relational ranking for stock prediction. ACM Transactions on Information Systems, 2019, 37 (2): 1- 30.
4	SAWHNEY R, AGARWAL S, WADHWA A, et al. Spatiotemporal hypergraph convolution network for stock movement forecasting[C]//Proceedings of the IEEE International Conference on Data Mining (ICDM). Washington D.C., USA: IEEE Press, 2020: 482-491.
5	ZHAO Y , DU H M , LIU Y , et al. Stock movement prediction based on bi-typed hybrid-relational market knowledge graph via dual attention networks. IEEE Transactions on Knowledge and Data Engineering, 2023, 35 (8): 8559- 8571. URL
6	HAMILTON J D , SUSMEL R . Autoregressive conditional heteroskedasticity and changes in regime. Journal of Econometrics, 1994, 64 (1/2): 307- 333. URL
7	ROTELA P J , SALOMON F L R . ARIMA: an applied time series forecasting model for the BOVESPA stock index. Applied Mathematics, 2014, 5 (21): 3383- 3391. doi: 10.4236/am.2014.521315
8	ZHANG L H, AGGARWAL C, QI G J. Stock price prediction via discovering multi-frequency trading patterns[C]//Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, USA: ACM Press, 2017: 1-7.
9	LI H, SHEN Y, ZHU Y. Stock price prediction using attention-based multi-input LSTM[EB/OL]. [2023-12-07]. https://proceedings.mlr.press/v95/li18c/li18c.pdf.
10	杨佳鹏. 基于注意力机制融合多特征的股票预测研究[D]. 北京: 北京交通大学, 2021.
	YANG J P. Research on stock prediction based on attention mechanism fusion of multiple features[D]. Beijing: Beijing Jiaotong University, 2021. (in Chinese)
11	LIN T, GUO T, ABERER K. Hybrid neural networks for learning the trend in time series[C]//Proceedings of the 26th International Joint Conference on Artificial Intelligence. Melbourne, Australia: International Joint Conferences on Artificial Intelligence Organization, 2017: 2273-2279.
12	CHEN C, ZHAO L, BIAN J, et al. Investment behaviors can tell what inside: exploring stock intrinsic properties for stock trend prediction[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York, USA: ACM Press, 2019: 2376-2384.
13	QIN Y, YANG Y. What you say and how you say it matters: predicting stock volatility using verbal and vocal cues[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Stroudsburg, USA: Association for Computational Linguistics, 2019: 1019-1038.
14	SEDINKINA M, BREITKOPF N, SCHVTZE H. Automatic domain adaptation outperforms manual domain adaptation for predicting financial outcomes[C]//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. Stroudsburg, USA: Association for Computational Linguistics, 2019: 1014-1034.
15	YING R, HE R N, CHEN K F, et al. Graph convolutional neural networks for Web-scale recommender systems[C]//Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. New York, USA: ACM Press, 2018: 974-983.
16	CHEN Y M, WEI Z Y, HUANG X J. Incorporating corporation relationship via graph convolutional neural networks for stock price prediction[C]//Proceedings of the 27th ACM International Conference on Information and Knowledge Management. New York, USA: ACM Press, 2018: 1655-1658.
17	LI W, BAO R H, HARIMOTO K, et al. Modeling the stock relation with graph network for overnight stock movement prediction[C]//Proceedings of the 29th International Joint Conference on Artificial Intelligence. Yokohama, Japan: International Joint Conferences on Artificial Intelligence Organization, 2020: 4541-4547.
18	KIM R, SO C H, JEONG M, et al. HATS: a hierarchical graph attention network for stock movement prediction[EB/OL]. [2023-12-07]. http://arxiv.org/abs/1908.07999.
19	HSU Y L , TSAI Y C , LI C T . FinGAT: financial graph attention networks for recommending top—profitable stocks. IEEE Transactions on Knowledge and Data Engineering, 2023, 35 (1): 469- 481. URL
20	YE J X, ZHAO J J, YE K J, et al. Multi-graph convolutional network for relationship-driven stock movement prediction[C]//Proceedings of the 25th International Conference on Pattern Recognition (ICPR). Washington D.C., USA: IEEE Press, 2021: 6702-6709.
21	MOODY J , SAFFELL M , LIAO Y S , et al. Reinforcement learning for trading systems and portfolios: immediate vs future rewards. Berlin, Germany: Springer, 1998.
22	MOODY J , WU L Z , LIAO Y S , et al. Performance functions and reinforcement learning for trading systems and portfolios. Journal of Forecasting, 1998, 17 (56): 441- 470. doi: 10.1002/(SICI)1099-131X(1998090)17:5/6<441::AID-FOR707>3.0.CO;2-#
23	MOODY J , SAFFELL M . Learning to trade via direct reinforcement. IEEE Transactions on Neural Networks, 2001, 12 (4): 875- 889. doi: 10.1109/72.935097
24	JIANG Z Y, XU D X, LIANG J J. A deep reinforcement learning framework for the financial portfolio management problem[EB/OL]. [2023-12-07]. https://arxiv.org/abs/1706.10059.
25	CHENG R, LI Q. Modeling the momentum spillover effect for stock prediction via attribute-driven graph attention networks[C]//Proceedings of the AAAI Conference on Artificial Intelligence. Palo Alto, USA: AAAI Press, 2021: 55-62.
26	KIPF T N, WELLING M. Semi-supervised classification with graph convolutional networks[EB/OL]. [2023-12-07]. http://arxiv.org/abs/1609.02907.
27	HU Z N, DONG Y X, WANG K S, et al. Heterogeneous graph Transformer[EB/OL]. [2023-12-07]. https://arxiv.org/abs/2003.01332.
28	VELIČKOVIĆ P, CUCURULL G, CASANOVA A, et al. Graph attention networks[EB/OL]. [2023-12-07]. http://arxiv.org/abs/1710.10903.

[1]	CHEN Depin, ZHAO Shen, JIAO Yiping, WANG Xiangxue, Lü Hong, XU Jun. Graph Construction on Whole Image Slides Based on Attention and Learnable Threshold [J]. Computer Engineering, 2025, 51(3): 229-240.
[2]	Zhengkang ZHANG, Dan YANG, Tiezheng NIE, Yue KOU. Self-Supervised Learning Based on Graph Structural Clustering for Disease Diagnosis Method [J]. Computer Engineering, 2024, 50(7): 360-371.
[3]	SONG Yanrui, ZHUANG Lei, XU Zexi, FENG Xu, MO Wenshuai. Reliable Service Function Chain Deployment Algorithm Based on Edge-Cloud Collaboration [J]. Computer Engineering, 2024, 50(12): 184-193.

Please choose a citation manager

Content to export