摘要
贷款风险分类是一个对借款人及其担保人的财务指标、非财务指标进行综合评价的过程。本文把贷款风险分类看作是一个模式识别问题,在此框架下,就统计模式识别领域中最新使用的神经网络方法、分类树法、K阶近邻方法以及支持向量机模型四种方法的建模思想、适用性进行比较,并给出有关结论。
The classification of credit risks is a comprehensive evaluation process about finance index and non - finance index of the borrower and guarantor. This paper regards the classification as a mode identification problem. Under this frame, the paper compares the modeling idea and applicability of four new statistic modes of nerve network, classification tree, K neighboring and support vector machine which are recently used in the field of statistic mode identification, and comes to relevant conclusions.
出处
《经济经纬》
CSSCI
北大核心
2008年第3期135-138,共4页
Economic Survey
关键词
模式识别
风险分类
适用性
mode identification
risk classification
applicability
