基于改进随机森林算法的上市公司信用风险实证分析
Empirical Analysis of Credit Risk of Listed Companies Based on Improved Random Forest Algorithm

近年来随着金融市场的不断发展，贷前识别风险企业、有效进行信贷风险控制越来越重要。本文主要研究企业信用风险评估的问题，通过合理的模型选择及模型优化，提升模型识别问题企业的能力。本文首先基于实际情况选择了合理的模型评估指标体系，通过优化后的随机森林算法，将特征选取与模型训练过程相结合，利用该模型以我国上市公司数据为例，进行了实证检验，并横向对比常见评估模型的数据表现，实验结果表明模型有较好的预测效果。
With the continuous development of the financial market in recent years, it has become more and more important to identify risky enterprises before lending and effectively control credit risks. This paper mainly studies the problem of enterprise credit risk assessment, and improves the ability of the model to identify problem enterprises through reasonable model selection and model optimization. This paper first selects a reasonable model evaluation index system based on the actual situation, and combines the feature selection and model training process through the optimized random forest algorithm. Comparing the data performance of common evaluation models, the experimental results show that the model has better prediction effect.