摘要
文章基于上市农业企业财务危机预测的3个常用机器学习模型,即贝叶斯、决策树、随机森林模型,从国泰安数据中心爬取2018—2020年上市农业公司财务数据以及近年来国家颁布的一系列涉农政策。通过数据预处理,缺失值的填充、多重共线性诊断、皮尔逊相关系数和SMOTE增加少数样本,最终通过归一化的数据进行机器学习模型的实验。结果显示决策树和随机森林模型具有较高的预测精度,其准确率达到了100%。此外文章还对影响上市农业公司的非传统财务指标国家政策进行分析,研究影响因素,得出结论,并通过提出解决措施,来确保上市农业公司能够实现财务稳定,走可持续发展之路。
Based on three commonly used machine learning models for the financial crisis prediction of listed agricultural enterprises,namely Bayes,Decision Tree and Random Forest Model,this paper crawls the financial data from 2018 to 2020 of listed agricultural companies on CSMAR and a series of agriculture-related policies issued by the state in recent years.Through data preprocessing,filling of missing values,multicollinearity diagnosis,Pearson correlation coefficient and adding several samples by SMOTE,the normalized data is finally brought into experiments based on machine learning model.The result finds that the Decision Tree and Random Forest Model show higher prediction accuracy up to 100%.In addition,this paper also analyzes the national policies of non-traditional financial indicators that affect listed agricultural companies,studies their influencing factors,draws conclusions,and puts forward solutions to ensure that listed agricultural companies can achieve financial stability and set forth towards sustainability.
作者
焦文豪
JIAO Wenhao(School of Accounting,Lanzhou University of Finance and Economics,Lanzhou 730020,China)
出处
《商业观察》
2025年第34期53-56,共4页
BUSINESS OBSERVATION
关键词
机器学习
财务危机
上市农业公司
machine learning
financial crisis
listed agricultural companies
