个人信用评估的加权组合预测模型  

Personal Credit Scoring Method Based on Weighted Combineing Model

作　　者：谢行恒[1] 陈玉芳[2] 

机构地区：[1]宁波工程学院经济与管理学院,浙江宁波315211 [2]杭州电子科技大学软件职业技术学院,浙江杭州310012

出　　处：《技术经济与管理研究》2008年第2期8-10,共3页Journal of Technical Economics & Management

摘　　要：目前单一的个人信用评估模型发展的很快,而精确性与稳定性却不能同时具备,且第二类误判率的能力弱,本文立足于用线性和非线性方法在个人信用评估中体现的优势,选择具有代表性的logistic回归和径向基函数神经网络方法,通过加权组合对个人信用进行预测评估,提高了评估模型的精确性与稳健性。With the rapid progress of the personal credit model, the accuracy and stability can not be improved at the same time. Based on the advantages of linear method with non- linear method in the personal credit scoring, chose typical logistic regression and RBF neural network to evaluate the personal credit through weighted combining model, which improve the accuracy and stability of the scoring model.

关 键 词：LOGISTIC回归 神经网络 组合预测 个人信用 

分 类 号：F224.9[经济管理—国民经济]