| 考虑宏观经济与企业间关联风险的中小企业破产预测 |
| Bankruptcy prediction for SMEs considering macroeconomy and inter-firm relational risks |
| 摘要点击 39 全文点击 0 投稿时间：2024-10-16 修订日期：2026-04-15 |
| 查看/发表评论 下载PDF阅读器 |
| 中文关键词 中小企业破产; 宏观经济指标; 企业关联风险; MHeCRK模型 |
| 英文关键词 SMEs bankruptcy; macroeconomic indicators; inter–firm relational risks; MHeCRK model |
| 基金项目 国家社会科学基金资助项目 |
| 投稿方向 金融系统建模与预测 |
| 作者 | 单位 | 邮编 | | 李杰 | 河北工业大学经济管理学院 | 300401 | | 宁尚卿* | 河北工业大学经济管理学院 | 300401 | | 李艳 | 河北工业大学经济管理学院 | |
|
| 中文摘要 |
| 中小企业破产预测是维护金融机构信贷安全和增强企业活力的重要因素. 相关研究多考虑企业微观风险而忽略宏观经济风险及企业关联风险. 本研究通过构建中小企业宏观经济指标丰富破产预测指标体系, 同时在异质图网络中利用多头与关系级注意力机制刻画中小企业间关联风险, 进一步与中小企业微观信息聚合建立破产预测模型 MHeCRK. 模型评估表明, 将宏观指标加入中小企业指标体系能够更加准确地评估企业经营状态, 基于图多头注意力机制的信息聚集网络则能够充分捕捉目标企业关联信息. 因此, 考虑中小企业自身风险的同时关注企业宏观风险, 并利用图注意力机制分析企业关联风险可以提高破产预测的准确率, 为中小企业相关决策提供更有效的参考和帮助. |
| 英文摘要 |
| Predicting bankruptcy for small and medium-sized enterprises (SMEs) is a crucial factor in maintaining the credit security of financial institutions and enhancing the vitality of businesses. Existing studies often focus on the micro-level risks of enterprises while neglecting macroeconomic risks and inter-firm relational risks.This study enriches the bankruptcy prediction indicator system for SMEs by incorporating macroeconomic indicators, while also modeling inter-enterprise relational risks using a multi-head and relation-level attention mechanism within a heterogeneous graph network. Furthermore, the aggregation of macroeconomic indicators with SMEs’ micro-level information forms the basis for the bankruptcy prediction model, MHeCRK. Model evaluations show that integrating macroeconomic indicators into the SMEs indicator system enables a more accurate assessment of an enterprise’s operational status. Additionally, the information aggregation network, based on the graph multi-head attention mechanism, can effectively capture the inter-firm relational information of the target enterprises. Therefore, by considering both the inherent risks of SMEs and their macroeconomic risks, and leveraging graph attention mechanisms to analyze inter-firm relational risks, the accuracy of bankruptcy prediction can be improved, providing more effective references and support for SMEs-related decision-making. |
| 关闭 |
|
|
|
|
|
