| 摘 要: 中小微企业在发展过程中需要信贷融资，然而部分企业的不良贷款会对金融稳定构成威胁。针对此问题，可以通过对企业相关财务数据进行分析，从源头上防范和降低信贷风险。本文采用熵权法、数据统计分析法、K均值聚类和BP神经网络等机器学习算法来给出相应的信贷策略，规避信贷风险。实验基于2020 年全国大学生数学建模竞赛赛题发布的公开数据集，利用MATLAB R2018a和Python 3.9等工具进行代码编写。测试结果表明，本文方法可以有效地对中小微企业的信贷风险进行评估并制定相应的信贷策略。 |
| 关键词: 风险评级模型；机器学习；熵权法；BP神经网络；K均值聚类 |
|
中图分类号: TP391
文献标识码: A
|
| 基金项目: 中国民航信息技术科研基地开放课题基金项目(CAAC-ITRB-201707)；河南省科技攻关项目(192102210283)；河南省高校重点科研项目(20A520040)；国家级大学生创新训练计划项目(202110485005). |
|
| Credit Risk Quantification and Decision Analysis based on Machine Learning |
|
WANG Zhaoxiang1, GE Lin1,2
|
( 1.School of Intelligent Engineering, Zhengzhou University of Aeronautics, Zhengzhou 450046, China ;
2.Information Technology Research Base of Civil Aviation Administration of China, Civil Aviation University of China, Tianjin 300300, China)
1113696260@qq.com; lingesnow@126.com
|
| Abstract: Medium, small and micro enterprises need credit financing in the development process, but non-performing loans of some enterprises pose a threat to financial stability. Aiming at this problem, an analysis of relevant financial data of the enterprise prevents and reduces credit risks from the source. This paper proposes corresponding credit strategies to avoid credit risks by using machine learning algorithms such as entropy weight method, data statistical analysis method, K-means clustering and BP neural network. The experiment is based on the public data set released by the 2020 National College Students Mathematical Modeling Contest, using MATLAB R2018a, Python 3.9 and other tools for code writing. Test results show that the proposed strategies can effectively evaluate the credit risk of medium, small and micro enterprises and formulate corresponding credit strategies. |
| Keywords: risk rating model; machine learning; entropy weight method; BP neural network; K-means clustering |
