摘要
研究目标:研究多源数据普适模型栈在信用评级中的构建与应用。研究方法:从我国在线信贷行业实际情况出发,提出一种基于多源数据的普适模型栈评分框架,该框架可以根据各个申请人不同的数据基础,自由选择纳入评分模型数据,生成子评分模型,然后再将子评分模型转换为常见的信用评分卡模型。研究发现:基于多源数据的普适模型栈评分框架不但灵活、普适,其评分有效性也比单个XGBoost信用评分模型更好。研究创新:将机器学习模型与传统评分卡模型进行了完美的融合。研究价值:解决了机器学习模型在信用风险管理中可解释性差的问题。
Research Objectives:General stack scoring framework based on the data from a variety of channels.Research Methods:Based on the actual situation of the online lending industry in China,this paper proposes a general stack scoring framework based on the data from a variety of channels.This framework can freely select the data of the score model according to the different data base of each lender,and then the sub-scoring model is transformed into a common credit scoring card model.Research Findings:The general stack scoring framework is not only flexible and general,but also has better effectiveness than a single XGBoost credit scoring model.Research Innovations:The machine learning model and the traditional scorecard model are perfectly integrated.Research Value:The problem of poor interpretability of machine learning in credit risk management is solved.
引文
[1]Agarwal S.,Hauswald R.B.H.,2008,The Choice between Arm's-Length and Relationship Debt:Evidence from Eloans[R],Federal Reserve Bank of Chicago Working Paper Series 2008-10.
[2]Altman E.L.,1968,Financial Ratios,Discriminant Analysis and the Prediction of Corporate Bankruptcy[J],Journal of Finance,23(4),589~609.
[3]Altaman E.I.,Marco G.,Varetto F.,1994,Corporate Distress Diagnosis:Comparisons Using Linear Discriminant Analysis and Neural Networks[J],Journal of Banking and Finance,18(3),505~529.
[4]Altman E.I.,Haldeman R.G.,Narayanan P.,1977,ZETA Analysis:A New Model to Identify Bankruptcy Risk of Corporations[J],Journal of Banking and Finance,1(1),29~54.
[5]Angelini P.,2000,Are Banks Risk Averse?Intraday Timing of Operations in the Interbank Market[J],Journal of Money Credit&Banking,32(1),54~73.
[6]Bachmann A.,Becker A.,Buerckner D.,Hilker M.,Kock F.,Lehmann M.,Tiburtius P.,Funk B.,2011,Online Peer-to-Peer Lending-A Literature Review[J],Journal of Internet Banking&Commerce,16(2),1~18.
[7]Beaver W.H.,1966,Financial Ratios As Predictors of Failure[J],Journal of Accounting Research,4,71~111.
[8]Herzenstein M.,Andrews R.L.,Dholakia U.M.,Lyandres E.,2008,The Democratization of Personal Consumer Loans?Determinants of Success in Online Peer-to-Peer Lending Communities[R],Boston University,School of Management,Research Paper,2008.
[9]Lin M.,2009,Peer-to-Peer Lending:An Empirical Study[R],AMCIS 2009Doctoral Consortium.
[10]Longstaff F.A.,Schwartz E.S.,1995,A Simple Approach to Valuing Risky Fixed and Floating Rate Debt[J],Journal of Finance,50(3),789~819.
[11]Messier W.F.,Hansen J.,1988,Inducing Rules for Expert System Development:An Example Using Default and Bankruptcy Data[J],Management Science,34(12),1403~1415.
[12]Odom M.D.,Sharda R.,1990,A Neural Network Models for Bankruptcy Prediction[R],1990IJCNN International Joint Conference on Neural Networks.
[13]Verstein A.,2011,The Misregulation of Person-to-Person Lending[J],UC Davis Law Review,45(2),445~529.
[14]李从刚、董中文、曹筱钰:《基于BP神经网络的P2P网贷市场信用风险评估》[J],《管理现代化》2015年第4期。
[15]李琦、曹国华:《基于Credit Risk+模型的互联网金融信用风险估计》[J],《统计与决策》2015年第19期。
[16]李太勇、王会军、吴江、张智林、唐常杰:《基于稀疏贝叶斯学习的个人信用评估》[J],《计算机应用》2013年第11期。
[17]李旭升、郭春香、郭耀煌:《扩展的树增强朴素贝叶斯网络信用评估模型》[J],《系统工程理论与实践》2008年第6期。
[18]沈沛龙、任若恩:《现代信用风险管理模型和方法的比较研究》[J],《经济科学》2002年第3期。
[19]王春峰、万海晖、张维:《商业银行信用风险评估及其实证研究》[J],《管理科学学报》1998年第1期。
[20]王锦虹:《基于逆向选择的互联网金融P2P模式风险防范研究》[J],《财经问题研究》2015年第5期。
[21]熊志斌:《基于非线性主成分分析的信用评估模型研究》[J],《数量经济技术经济研究》2013年第10期。
[22]杨胜刚、朱琦、成程:《个人信用评估组合模型的构建:基于决策树-神经网络的研究》[J],《金融论坛》2013年第2期。
[23]于晓虹、楼文高:《基于随机森林的P2P网贷信用风险评价、预警与实证研究》[J],《金融理论与实践》2016年第2期。
[24]张玲:《财务危机预警分析判别模型及其应用》[J],《数量经济技术经济研究》2000年第3期。