AI-based risk assessment and early warning method for futures trading

Chang Fei; Wang Junjie; Hu Yi; Qiao Yan; Weng Zhenhao; Guo Jinsheng; Liu Yibo; Li Meng

doi:10.11959/j.issn.2096-109x.2026005

您当前的位置：

首页 >

文章列表页 >

AI-based risk assessment and early warning method for futures trading

Papers | 更新时间：2026-05-26

- AI-based risk assessment and early warning method for futures trading
- Chinese Journal of Network and Information Security Vol. 12, Issue 1, Pages: 151-163(2026)
- 作者机构：
  
  1.徽商期货有限责任公司，安徽合肥 230002
  2.合肥工业大学计算机与信息学院，安徽合肥 230601
  3.合肥工业大学大数据知识工程教育部重点实验室，安徽合肥 230601
- 作者简介：
- 基金信息：
  
  The 18th CFA Joint Research Program;Research Project of China Securities Regulatory Commission's Technology Regulation(HSQHJSZX-2025-01);The National Natural Science Foundation of China(62572168;62372149;U23A20303);The Natural Science Foundation of Anhui Province(2508085MF151);Open Project of the Key Laboratory of Knowledge Engineering with Big Data, Ministry of Education(BigKEOpen2025-04)
- DOI：10.11959/j.issn.2096-109x.2026005
  CLC： F830.9
- Received：23 June 2025，
  
  Revised：2025-12-11，
  
  Published：25 February 2026
- 稿件说明：
移动端阅览
常飞, 王俊杰, 胡毅, 等. 基于人工智能的期货交易风险评估与预警方法[J]. 网络与信息安全学报, 2026, 12(1): 151-163.

Chang Fei, Wang Junjie, Hu Yi, et al. AI-based risk assessment and early warning method for futures trading[J]. Chinese Journal of Network and Information Security, 2026, 12(1): 151-163.
常飞, 王俊杰, 胡毅, 等. 基于人工智能的期货交易风险评估与预警方法[J]. 网络与信息安全学报, 2026, 12(1): 151-163. DOI： 10.11959/j.issn.2096-109x.2026005.

Chang Fei, Wang Junjie, Hu Yi, et al. AI-based risk assessment and early warning method for futures trading[J]. Chinese Journal of Network and Information Security, 2026, 12(1): 151-163. DOI： 10.11959/j.issn.2096-109x.2026005.

摘要

随着期货市场交易复杂性和系统性风险的不断攀升，传统规则驱动的监管手段已难以有效识别高风险的交易行为。为此，面向期货行业交易风险预警与市场监管的实际需求，提出一种基于人工智能的期货交易风险评估与预警方法（Fut-GAT-LSTM）。首先，结合期货交易特点，筛选可能引发市场动荡的多维度交易特征，针对不同的特征类型设计了不同的交易特征嵌入方法，并将嵌入的特征数据转换为时间序列图结构；随后，通过图注意力网络（GAT）建模交易图内部的空间依赖关系，利用双向长短期记忆网络（BiLSTM）提取交易行为随时间演化的动态特征，并基于交叉熵损失进行有监督训练；最后，通过训练后的模型计算每个交易节点的风险评分，并利用动态阈值进行交易风险的预警。使用某大型期货公司提供的两组数据开展实验，结果显示所提模型在AUC值、F1-score等关键指标上均显著优于传统分类模型和神经网络基线方法，在强平风险和市场动荡风险检测任务中均表现出优异的性能。该方法为构建精准、高效、可扩展的智能期货市场监管系统提供了新思路，具有良好的实际应用价值与推广前景。

Abstract

With the increasing complexity of futures trading and rising systemic risks

traditional rule-based regulatory approaches can no longer effectively identify high-risk trading activities. In this paper

an artificial intelligence-based method for risk assessment and early warning in the context of futures trading risk monitoring and supervision

named Fut-GAT-LSTM

was proposed. Firstly

based on the characteristics of futures trading

multidimensional trading features likely to trigger market turbulence were selected. Different embedding methods were designed for different types of trading features

and the embedded feature data were transformed into a time-series graph structure. Subsequently

a graph attention network (GAT) was employed to model the spatial dependencies within the trading graph

while a bidirectional long short-term memory (BiLSTM) network was used to extract the temporal dynamics of trading behaviors evolved over time. The model was trained in a supervised manner based on cross-entropy loss. Finally

risk scores for each trading node were calculated by the trained model

and dynamic thresholds were utilized for risk early warning. Experiments conducted on two experimental datasets from a large futures company demonstrated that the proposed model significantly outperformed traditional classification models and baseline neural network approaches in key metrics such as AUC and F1-score

achieving superior performance in both forced liquidation risk detection and market turbulence risk detection tasks. This method offers a novel approach for building accurate

efficient

and scalable intelligent regulatory systems for the futures market

possessing strong practical value and application potential.

关键词

Keywords

references

熊熊 , 袁海亮 , 张维 , 等 . 程序化交易及其风险分析 [J ] . 电子科技大学学报(社科版) , 2011 , 13 ( 3 ): 32 - 39 .

Xiong X , Yuan H L , Zhang W , et al . Program trading and its risk analysis [J ] . Journal of University of Electronic Science and Technology of China (Social Sciences Edition) , 2011 , 13 ( 3 ): 32 - 39 .

朱里 . 美国高频交易监管机制研究 [D ] . 上海：华东政法大学 , 2020 .

Zhu L . Research on the supervision mechanism of high frequency trading in the united states [D ] . Shanghai : East China University of Political Science and Law , 2020 .

陈汇 . 幌骗犯罪论 [D ] . 上海 : 华东政法大学 , 2020 .

Chen H . Theory of spoofing [D ] . Shanghai : East China University of Political Science and Law , 2020 .

胡乔 . 证券投资基金与股市波动性——基于中国的理论与实证 [D ] . 南昌：江西财经大学 , 2009 .

Hu Q . Security investment funds and volatility of stock market: theory and evidence based on china [D ] . Nanchang : Jiangxi University of Finance and Economics , 2009 .

方建珍 . 信用风险转移机制研究 [D ] . 武汉 : 武汉理工大学 , 2025 .

Fang J Z . A research on credit risk transfer mechanism [D ] . Wuhan : Wuhan University of Technology , 2025 .

王欣 , 尹留志 , 方兆本 . 异常交易行为的甄别研究 [J ] . 数理统计与管理 , 2009 , 28 ( 4 ): 671 - 677 .

Wang X , Yin L Z , Fang Z B . Research of detecting outlier transaction behavior [J ] . Journal of Applied Statistics and Management , 2009 , 28 ( 4 ): 671 - 677 .

金升平 , 刘钊 . 银行异常交易检测方法研究 [J ] . 武汉金融 , 2018 , ( 2 ): 61 - 66 .

Jin S P , Liu Z . Research on anomaly detection methods for bank transactions [J ] . Wuhan Finance Monthly , 2018 ( 2 ): 61 - 66 .

张启宏 . 基于人工智能的金融监管信息系统 [J ] . 现代计算机(专业版) , 2002 ,( 6 ): 49 - 51+60 .

Zhang Q H . A Banking supervision information system based on artificial intelligence [J ] . Modern Computer (Professional Edition) , 2002 , ( 6 ): 49-51 + 60 .

Liu Z Q , Chen C C , Yang X X , et al . Heterogeneous graph neural networks for malicious account detection [C ] // Proceedings of the 27th ACM International Conference on Information and Knowledge Management . New York : ACM , 2018 : 2077 - 2085 .

Quah J T S , Sriganesh M . Real-time credit card fraud detection using computational intelligence [J ] . Expert Systems with Applications , 2008 , 35 ( 4 ): 1721 - 1732 .

Yu W C , Cheng W , Aggarwal C C , et al . NetWalk: a flexible deep embedding approach for anomaly detection in dynamic networks [C ] // Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery Data Mining . New York : ACM , 2018 : 2672 - 2681 .

Zheng L , Li Z P , Li J , et al . AddGraph: anomaly detection in dynamic graph using attention-based temporal GCN [C ] // Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence . International Joint Conferences on Artificial Intelligence Organization , 2019 : 4419 - 4425 .

Liu Y X , Pan S R , Wang Y G , et al . Anomaly detection in dynamic graphs via transformer [J ] . IEEE Transactions on Knowledge and Data Engineering , 2023 , 35 ( 12 ): 12081 - 12094 .

Michalak K , Korczak J . Graph mining approach to suspicious transaction detection [C ] // Proceedings of the 2011 Federated Conference on Computer Science and Information Systems (FedCSIS) . Piscataway : IEEE Press , 2011 : 69 - 75 .

Ding K Z , Zhou Q H , Tong H H , et al . Few-shot network anomaly detection via cross-network meta-learning [C ] // Proceedings of the Web Conference 2021 . New York : ACM , 2021 : 2448 - 2456 .

Aggarwal C C , Zhao Y C , Yu P S . Outlier detection in graph streams [C ] // Proceedings of the 2011 IEEE 27th International Conference on Data Engineering . Piscataway : IEEE Press , 2011 : 399 - 409 .

张建刚 . 期货市场自律监管体系比较与借鉴 [J ] . 武汉理工大学学报(社会科学版) , 2005 , 18 ( 5 ): 694 - 698 .

Zhang J G . International comparison of futures market self-regulation system and its enlightenment to china [J ] . Journal of Wuhan University of Technology(Social Sciences Edition) , 2005 , 18 ( 5 ): 694 - 698 .

尚春梅 , 刘继曾 . 期货交易的特征与功能 [J ] . 山西财经大学学报 , 1994 ( 5 ): 54 - 55,57 .

Shang C M , Liu J Z . Characteristics and functions of futures trading [J ] . Journal of Shanxi University of Finance and Economics , 1994 ( 5 ): 54 - 55,57 .

Zhang Q Y , Zhang Y F , Yao X X , et al . A dynamic attributes-driven graph attention network modeling on behavioral finance for stock prediction [J ] . ACM Transactions on Knowledge Discovery from Data , 2024 , 18 ( 1 ): 1 - 29 .

Siami-Namini S , Tavakoli N , Namin A S . The performance of LSTM and BiLSTM in forecasting time series [C ] // Proceedings of the 2019 IEEE International Conference on Big Data (Big Data) . Piscataway : IEEE Press , 2020 : 3285 - 3292 .

Zong B , Song Q , Min M R , et al . Deep autoencoding Gaussian mixture model for unsupervised anomaly detection [C ] // Proceedings of the International Conference on Learning Representations . 2018 .

Park D , Hoshi Y , Kemp C C . A multimodal anomaly detector for robot-assisted feeding using an LSTM-based variational autoencoder [J ] . IEEE Robotics and Automation Letters , 2018 , 3 ( 3 ): 1544 - 1551 .

Du B W , Sun X X , Ye J C , et al . GAN-based anomaly detection for multivariate time series using polluted training set [J ] . IEEE Transactions on Knowledge and Data Engineering , 2023 , 35 ( 12 ): 12208 - 12219 .

Tuli S , Casale G , Jennings N R . TranAD: deep transformer networks for anomaly detection in multivariate time series data [C ] // Proceedings of the VLDB Endowment , 2022 , 15 ( 6 ): 1201 - 1214 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Technology and practice of intelligent governance for financial data security

Adaptive selection method of desensitization algorithm based on privacy risk assessment

Preliminary study on the reform of machine learning teaching

Research on network risk assessment based on attack graph of expected benefits-rate

Information security vulnerability scoring model for intelligent vehicles

Related Author

Hongfeng CHAI

Yinan JIN

Hanbing ZHANG

Yihao CUI

Dan CHEN

Xu SUN

Fei ZHANG

Zhixin LI

Related Institution

School of Computer Science, Fudan University

Institute of Financial Technology, Fudan University

School of Financial Technology, Fudan University

China UnionPay Co., Ltd

Huawei Technologies Co., Ltd

⁰