一种面向非均衡样本的企业金融风险预测方法  

Enterprise Financial Risk Forecasting Method Based on Unbalanced Samples

作　　者：仇实 高影繁[1] 姚长青[1] 刘志辉[1] 李佳星 QIU Shi;GAO Yingfan;YAO Changqing;LIU Zhihui;LI Jiaxing(Institute of Scientific and Technical Information of China,Beijing 100038;Bank of Hebei,Shijiazhuang 050011)

机构地区：[1]中国科学技术信息研究所,北京100038 [2]河北银行,河北石家庄050011

出　　处：《中国科技资源导刊》2021年第5期11-17,共7页China Science & Technology Resources Review

基　　金：中国科学技术信息研究所重点工作项目“上市公司年报数据库建设及服务系统研发”(ZD2021-10)。

摘　　要：在企业所面临的众多风险中,企业金融风险表现尤为突出,而在大数据环境下,严重的数据不均衡成为横亘在企业金融风险分析面前的一道鸿沟。本文针对企业竞争情报分析中的样本不均衡问题,以金融企业信贷风险预测为切入点,提出一种面向非平衡样本的企业风险识别方法。该方法采用人工智能分析领域中的特征选择、非均衡样本平衡处理和集成学习等智能分析手段,为大数据环境下企业竞争情报中的企业风险识别问题提供解决思路。Among the many risks faced by enterprises,corporate financial risks bear the brunt.In the context of big data,serious data imbalance has become a chasm in front of the analysis of corporate financial risks.Aiming at the problem of sample imbalance in competitive intelligence analysis of enterprises,this paper takes credit risk prediction of financial enterprises as the breakthrough point,and puts forward an enterprise risk identification method for unbalanced samples.This method adopts a variety of intelligent analysis methods in the field of artificial intelligence analysis,such as feature selection,unbalanced sample balance processing and ensemble learning,to provide solutions to the problem of enterprise risk identification in enterprise competitive intelligence under the environment of big data.

关 键 词：非平衡样本 金融企业 信贷风险 Catboost EasyEnsemble 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]