征信
徵信
정신
CREDIT REFERENCE
2015年
2期
11-14
,共4页
个人信用评价%logistic模型%不平衡数据%决策分析
箇人信用評價%logistic模型%不平衡數據%決策分析
개인신용평개%logistic모형%불평형수거%결책분석
personal credit evaluation%logistic model%imbalanced data%decision analysis
Logistic模型是个人信用评价分析时常用的模型。基于具体的个人信贷数据,利用Random Oversampling方法克服了数据不平衡性带来的问题,通过共线性诊断后,建立了标准化的logistic回归模型。根据标准化回归系数,计算出决策系数,得出个人特征对信用评价影响大小的顺序。
Logistic模型是箇人信用評價分析時常用的模型。基于具體的箇人信貸數據,利用Random Oversampling方法剋服瞭數據不平衡性帶來的問題,通過共線性診斷後,建立瞭標準化的logistic迴歸模型。根據標準化迴歸繫數,計算齣決策繫數,得齣箇人特徵對信用評價影響大小的順序。
Logistic모형시개인신용평개분석시상용적모형。기우구체적개인신대수거,이용Random Oversampling방법극복료수거불평형성대래적문제,통과공선성진단후,건립료표준화적logistic회귀모형。근거표준화회귀계수,계산출결책계수,득출개인특정대신용평개영향대소적순서。
Logistic model is commonly used in the analysis of personal credit evaluation. Based on personal credit data, the paper first of all overcame the problems resulting from imbalanced data by using Random Oversampling method, and then established a standardized logistic regression model through co-linearity diagnostics. Decision coefficient was calculated according to standardized regression coefficients, then finding out the influence order of personal characteristics on credit evaluation.