Evolution Mechanism and Path Analysis of Industrial Chain Risks Based on Large Language Models and Eventic Graph
Deng Zhe1,2,3,4, Ma Jianxia1,2,3,4
1. Northwest Institute of Eco-Environment and Resources,Chinese Academy of Sciences,Lanzhou 730000,China; 2. Key Laboratory of Ecological Safety and Sustainable Development in Arid Lands,Northwest Institute of Eco-Environment and Resources,Chinese Academy of Sciences,Lanzhou 730000; 3. Key Laboratory of Knowledge Computation and Decision Intelligence of Gansu Province,Lanzhou 730000,China; 4. Department of Information Resources Management,School of Economics and Management,University of Chinese Academy of Sciences,Beijing 100190,China
Abstract：Against the backdrop of globalization and increasing complexity in industrial chains,traditional industrial analysis methods struggle to reveal the dynamic evolution patterns of risks.To address this,this paper proposes an analytical framework that integrates large language models(LLMs)and eventic graph for the dynamic identification and evolutionary path analysis of industrial chain risks.The framework first designs prompts based on a risk event ontology to guide LLMs in risk news discrimination and element extraction.It then employs BERTopic for event generalization and constructs an abstract eventic graph.Finally,complex network analysis is integrated to identify key nodes and evolutionary chains.An empirical study focusing on the industrial chain of new energy vehicle demonstrates that the proposed LLM-based methods for risk news discrimination and event element extraction outperform traditional approaches.A typical case study on“tariff increases”reveals a chain-like evolutionary mechanism:“external shock → cost transmission → market response → multi-stakeholder response.”In conclusion,the LLM-driven eventic graph method effectively delineates the evolutionary paths of industrial chain risks,providing insights for explaining risk transmission mechanisms and informing risk governance strategies.
邓喆, 马建霞. 基于大模型与事理图谱的产业链风险演化机制与路径分析[J]. 中国科技论坛, 2026(6): 87-97.
Deng Zhe, Ma Jianxia. Evolution Mechanism and Path Analysis of Industrial Chain Risks Based on Large Language Models and Eventic Graph. , 2026(6): 87-97.
[1] 汤惠民,袁媛,李家琳,等.数字产业链背景下的产业链安全风险评估[J].中国信息化,2023(10):38-41. [2] 石建勋,卢丹宁.着力提升产业链供应链韧性和安全水平研究[J].财经问题研究,2023(2):3-13. [3] 郭靖怡,王学昭,陈小莉.基于专利文本中产品关联关系的产业技术链构建与实证研究:以锂离子电池产业为例[J].图书情报工作,2023,67(5):108-118. [4] 张静,黄文锋,吴春江,等.大模型增强知识图谱的构建与推理研究综述[J].计算机科学与探索,2025,19(11):2855-2872. [5] 刘家国,孔玉丹,周欢,等.供应链风险管理的物理-事理-人理方法研究[J].系统工程学报,2018,33(3):298-307. [6] 马林.基于SCOR模型的供应链风险识别、评估与一体化管理研究[D].杭州:浙江大学,2005. [7] 陈国华,张根保,任显林,等.基于故障树分析法的供应链可靠性诊断方法及仿真研究[J].计算机集成制造系统,2009,15(10):2034-2038,2049. [8] 沈玉燕,钱言.生鲜冷链供应链风险扩散收敛模型构建与评估[J].商业经济研究,2021(22):46-49. [9] ADHITYA A,SRINIVASAN R,KARIMI I A.Supply chain risk identification using a HAZOP-based approach[J].AIChE Journal,2009,55(6):1447-1463. [10] ORDIBAZAR H A,HUSSAIN K O,CHAKRABORTTY K R,et al.Artificial intelligence applications for supply chain risk management considering interconnectivity,external events exposures and transparency:A systematic literature review[J].Modern Supply Chain Research and Applications,2025,7(2):148-179. [11] KOSASIH E E,MARGAROLI F,GELLI S,et al.Towards knowledge graph reasoning for supply chain risk management using graph neural networks[J].International Journal of Production Research,2024,62(15):5596-5612. [12] GANESH A D,KALPANA P.Supply chain risk identification:A real-time data-mining approach[J].Industrial Management & Data Systems,2022,122(5):1333-1354. [13] CHU C Y,PARK K,KREMER G E.A global supply chain risk management framework:An application of text mining to identify region-specific supply chain risks[J].Advanced Engineering Informatics,2020,45:101053. [14] GUAN S,CHENG X,BAI L,et al.What is event knowledge graph:A survey[J].IEEE Transactions on Knowledge and Data Engineering,2023,35(7):7569-7589. [15] CAO Y,LAN Y,ZHAI F,et al.5W1H extraction with large language models[C]//2024 International Joint Conference on Neural Networks(IJCNN).Piscataway:IEEE,2024:1-8. [16] WEI X,CUI X,CHENG N,et al.ChatIE:Zero-shot information extraction via chatting with ChatGPT[J].ArXiv Preprint ArXiv:2302.10205,2023. [17] 鲍彤,章成志.ChatGPT中文信息抽取能力测评:以三种典型的抽取任务为例[J].数据分析与知识发现,2023,7(9):1-11. [18] 刘宗田,黄美丽,周文,等.面向事件的本体研究[J].计算机科学,2009,36(11):189-192,199. [19] 徐雷,潘珺.事件表示方式及其语义表示模型研究[J].情报杂志,2019,38(6):159-167. [20] JEINLEE1991.jeinlee1991/chinese-llm-benchmark[CP/OL].(2026-04-12)[2026-04-12].https://github.com/jeinlee1991/chinese-llm-benchmark. [21] 宋新平,刘馥宁,申真,等.大数据下企业供应链风险管理与竞争情报融合模型构建:以华为公司为例[J].情报杂志,2024,43(6):185-192,176. [22] TANG W,BU H,ZUO Y,et al.Unlocking the power of the topic content in news headlines:BERTopic for predicting Chinese corporate bond defaults[J].Finance Research Letters,2024,62:105062. [23] 胡凯茜,李欣,王龙腾.基于BERTopic模型的网络暴力事件衍生舆情探测[J].情报杂志,2024,43(7):146-153.
2026, 
