电子技术
電子技術
전자기술
ELECTRONIC TECHNOLOGY
2013年
9期
26-28
,共3页
违约概率%信用风险%Logistic%主成分分析
違約概率%信用風險%Logistic%主成分分析
위약개솔%신용풍험%Logistic%주성분분석
PD%credit risk%Logistic%principal component analysis
目前商业银行利润的主体主要来自信贷业务。因此，信贷风险一直是商业银行最为关注的风险之一，如何对信用风险评估是商业银行业务中很关注的问题。与传统的信用风险评估方法需要更多人力参与相比，本文提出一种直接从银行客户交易数据中抽取所关注的交易行为，然后在此基础上利用Logistic回归方法建立模型的方法，来预测反映了客户贷款偿还能力的违约概率（PD）。
目前商業銀行利潤的主體主要來自信貸業務。因此，信貸風險一直是商業銀行最為關註的風險之一，如何對信用風險評估是商業銀行業務中很關註的問題。與傳統的信用風險評估方法需要更多人力參與相比，本文提齣一種直接從銀行客戶交易數據中抽取所關註的交易行為，然後在此基礎上利用Logistic迴歸方法建立模型的方法，來預測反映瞭客戶貸款償還能力的違約概率（PD）。
목전상업은행리윤적주체주요래자신대업무。인차，신대풍험일직시상업은행최위관주적풍험지일，여하대신용풍험평고시상업은행업무중흔관주적문제。여전통적신용풍험평고방법수요경다인력삼여상비，본문제출일충직접종은행객호교역수거중추취소관주적교역행위，연후재차기출상이용Logistic회귀방법건립모형적방법，래예측반영료객호대관상환능력적위약개솔（PD）。
The main body of current commercial bank profits is mainly from the credit business. Therefore, the credit risk has been one of the most concerned risks of commercial banks, and how to assess the credit risk is a concern in commercial banking business. Compared with the traditional credit risk assessment methods which require more labor participation, this paper proposes a new method which is based on the data directly extracted from bank customer transaction database and uses Logistic regression method to model, so as to forecasts customer's PD reflecting the customer probability of default loan repayment ability.
