个人小额信贷风险影响因素研究  

Research on Microfinance Credit Risk Assessment

作　　者：李好奇 林华珍[2] 黄福 LI Hao-qi;LIN Hua-zhen;HUANG Fu(School of Mathematics and Statistics,Yangtze Normal University,Chongqing 408100,China;Center of Statistical Research,School of Statistics,Southwestern University of Finance and Economics,Chengdu 611130,China)

机构地区：[1]长江师范学院数学与统计学院,重庆408100 [2]西南财经大学统计研究中心,统计学院,四川成都611130

出　　处：《数理统计与管理》2022年第6期1105-1115,共11页Journal of Applied Statistics and Management

基　　金：国家自然科学基金(11571282,11829101);重庆市自然科学基金(cstc2019jcyj-msxmX0709);重庆市教委科学技术研究项目(KJQN201901436)。

摘　　要：随着互联网消费金融兴起,小额信贷慢慢渗入到经济生活中。然而,金融机构在获得消费信贷利润的同时,持续攀升的不良贷款率也在逼着这些信贷公司不断提高风险评估水平和控制能力。于是,个人信用风险的识别和信贷风险管理成为热门话题。本文基于一家互联网金融公司的个人信贷数据,利用未知连接函数的广义可加模型(GAMUL),分析各个特征变量如何影响个人信用风险。本文首先使用非参独立扫描(NIS)方法选择影响显著的变量,考虑了自变量与因变量之间的非线性关系,再使用GAMUL对数据建立模型。无论是变量选择还是模型分析,NIS-GAMUL都没有对自变量和响应变量进行过多的人为主观假设,完全由数据驱动,使整个模型分析具有较强的适应性和灵活性。As the rise of Internet consumer finance,microfinance gradually penetrated into all aspects of our lives.However,while financial institutions gain profits from consumer credit,the rising nonperforming loan ratio also forces these credit companies to continuously improve their risk assessment level and controlling ability.Therefore,the identification of personal credit risk and credit risk management has become a hot topic.In this paper,we analyze how each characteristic variable affects personal credit risk by the nonparametric additive model with unknown link function(GAMUL) based on the personal credit data of a internet finance company.In this paper,the non-parametric independent scanning(NIS)method is used to select variables,considering the non-linear relationship between independent variables and dependent variables,then,GAMUL was used to model the data.In terms of both variable selection and model building,NIS-GAMUL does not make too many artificial and subjective assumptions on independent variables and response variables.It is completely driven by data,which make the whole model highly adaptive and flexible.

关 键 词：小额货款 按期还款率 未知连接函数 广义可加模型 

分 类 号：O212[理学—概率论与数理统计] O212.1[理学—数学]