基于主成分分析和logistic回归的中小微企业信贷决策

Creditdecisions of small,medium and micro enterprises based on principal component analysis and logistic regression

中小微企业由于自身经营规模、持续经营时间等因素的影响,相较于大型企业而言,其信贷风险较高,因此往往会陷入“融资难”的困境中。高信贷风险已成为银行发展中小微企业信贷业务的掣肘,如何对中小微企业的信贷风险进行预测是发展中小微企业信贷业务的关键。首先根据中小微企业进销项发票信息进行特征提取和降维处理,再利用logistic回归计算企业的违约概率,最后通过目标规划进行信贷决策。结果表明,该模型所预测的信贷风险对信贷决策起着支撑作用,并为银行的信贷决策业务提供了一种有效方法。

Small,medium and micro enterprises have a higher credit risk than large enterprises due to their own business scale and duration of operation,so they often fall into the dilemma of“financing difficulties”.Higher credit risk has become a constraint for banks to conduct credit business for small,medium and micro enterprises.How to predict the credit risk of small,medium and micro enterprises and decision-making is the key to the development of credit business for small,medium and micro enterprises.Perform feature extraction and dimensionality reduction processing based on the invoice information of small,medium and micro enterprises'purchase and sale items.Logistic regression is used to calculate the default probability of the enterprise.Finally,credit decisions are made through target planning.The results show that the credit risk predicted by the model supports credit decision-making and provides an effective method for the bank's credit decision-making business.