基于DeepSeek大语言模型的企业财务危机预警研究  

Research on Enterprise Financial Crisis Early Warning Based on DeepSeek

作　　者：陈昌华[1,2] 伍华骅 熊自玉 兰杰 李佳怡 李波 章超[5] CHEN Chang-hua;WU Hua-hua;XIONG Zi-yu;LAN Jie;LI Jia-yi;LI Bo;ZHANG Chao(School of Management,Xihua University,Chengdu,Sichuan,610039,China;Institute of International Economics and Management,Xihua University,Chengdu,Sichuan,610039,China;Financial Management Office,China National Tobacco Corporation Sichuan Provincial Company,Chengdu,Sichuan,610017,China;School of Computer and Software Engineering,Xihua University,Chengdu,Sichuan,610039,China;Intelligent Policing Key Laboratory of Sichuan Province,Sichuan Police College,Luzhou,Sichuan,646000)

机构地区：[1]西华大学管理学院,四川成都610039 [2]西华大学国际经济与管理研究院,四川成都610039 [3]中国烟草总公司四川省公司财务管理处,四川成都610017 [4]西华大学计算机与软件工程学院,四川成都610039 [5]四川警察学院四川省智能警务重点实验室,四川泸州646000

出　　处：《西华大学学报(哲学社会科学版)》2025年第3期50-58,共9页Journal of Xihua University(Philosophy & Social Sciences)

基　　金：中国烟草总公司四川省公司科技项目(项目编号:SCYC202542;SCYC202325);四川省智能警务重点实验室资助项目(项目编号:ZNJW2022KFMS004)。

摘　　要：在全球经济政策环境多变的VUCA时代,企业财务危机呈现出非线性传导与多源风险耦合的新特征,传统预警模型面临失效风险。为此,文章利用DeepSeek强大的语言理解和文本处理能力,提出一种基于“数据-文本-分类”转换路径的企业财务危机预警方法,该方法将包含Z值和13个财务指标的结构化财务数据转化为流畅的企业资料描述,用于微调预训练模型性能,实现财务危机准确识别。文章以147家上市食品制造企业2002—2023年财务数据为样本,研究结果显示:基于DeepSeek生成文本微调后的预训练模型在准确率、F1宏平均、召回率宏平均上均超过95%,显著优于随机森林和DeepFM等传统模型。研究结果表明:DeepSeek大模型用于企业财务危机预警是有效的,可为AI重塑财务提供理论支撑和实践参考。In the VUCA era characterized by volatile global economic policies,corporate financial crises exhibit new features of nonlinear transmission and multi-source risk coupling,rendering traditional early warning models obsolete.To address this,we pro-pose a financial crisis early warning method based on the"data-text-classification"transformation path,leveraging the powerful lan-guage understanding and text processing capabilities of DeepSeek.This method converts structured financial data containing Z-score and 13 financial indicators into smooth enterprise data description,which is used to fine-tune the performance of pre-training model and realize accurate identification of financial crisis.Using financial data from 147 publicly listed food manufacturing enterprises spanning the years 2002 to 2023,experimental results show that the pre-trained model fine-tuned with DeepSeek-generated text achieves over 95%accuracy,macro-F1 score,and macro-average recall,significantly outperforming traditional models such as Random Forest and DeepFM.This study highlights the effectiveness of the DeepSeek model for enterprise financial crisis early warning,offering both the-oretical contributions and practical insights for advancing AI-driven financial risk management.

关 键 词：财务危机 危机预警 预训练模型 DeepSeek 

分 类 号：F275[经济管理—企业管理]