基于TEI@I方法论的借款人信用品质转移概率计算及失联概率预测应用

管理评论 ›› 2020, Vol. 32 ›› Issue (7): 267-279.

基于TEI@I方法论的借款人信用品质转移概率计算及失联概率预测应用

庞素琳^1,2, 侯鲜艳^1,2

1. 暨南大学管理学院金融工程研究所/应急管理学院, 广州 510632;
2. 广东省公共网络安全风险评价与预警应急技术研究中心, 广州 510632

收稿日期:2019-09-05 出版日期:2020-07-28 发布日期:2020-08-08
作者简介:庞素琳,暨南大学管理学院金融工程研究所/应急管理学院教授,博士生导师;侯鲜艳,暨南大学管理学院金融工程研究所/应急管理学院硕士研究生。
基金资助:
国家自然科学基金项目（91646112）。

Calculating Method of Borrower's Credit Quality Transfer Probability and Application in Forecasting Loss-of-Connection Probability Based on TEI@I Methodology

Pang Sulin^1,2, Hou Xianyan^1,2

1. Institute of Finance Engineering in School of Management/School of Emergency Management, Jinan University, Guangzhou 510632;
2. Guangdong Emergency Technology Research Center of Risk Evaluation and Pre-warning on Public Network Security, Guangzhou 510632

Received:2019-09-05 Online:2020-07-28 Published:2020-08-08

摘要/Abstract

摘要： 本文基于TEI@I方法论的理论框架，首次提出研究借款人信用品质转移概率计算及失联概率预测应用。文本挖掘用来处理借款人信用品质的不稳定性，根据信用等级和信用评分划分等级，模糊分类信用品质等级，并定义了信用品质等级转移概率。条件概率计量模型推广到信用等级条件概率和信用评分条件概率，利用全概率计算方法得出信用品质条件概率计算公式。借鉴标准普尔信用转移矩阵的计算方法，同时再运用马尔科夫C-K方程对借款人信用品质等级转移概率进行预测，以此联合研究借款人失联的概率。研究结果表明：信用品质等级越高，借款人按时还款的概率越高，信用品质等级越低，借款人坏账的概率越高；越靠近"失联"等级的借款人，其失联可能性越大；随着借款期限的延长，信用品质等级保持原始等级的可能性不大，转向下一等级的转移概率会逐渐上升。

关键词: TEI@I方法论, 信用等级与信用评分, 信用品质转移概率, 失联概率预测, C-K方程

Abstract: Based on the theoretical framework of TEI@I methodology, for the first time, we study the calculation method of the probability of borrower's credit quality and its forecasting probability prediction application. In this paper, text mining is used to deal with the instability of borrower's credit quality, credit grade and credit score are used for classification, credit quality grade is classified by fuzzy classification, and the conversion probability of credit quality is defined. The conditional probability measurement model is extended to the conditional probability of credit grade and the conditional probability of credit score. The calculation formula of conditional probability of credit quality is obtained by using the method of total probability calculation. Referring to the calculation method of S&P credit transfer matrix, and using Markov C-K equation to predict the transfer probability of borrower's credit quality grade, this paper jointly studies the probability of borrower's loss of connection. The results show that the higher the credit quality grade is, the higher the probability of timely repayment is, the lower the credit quality grade is, and the higher the probability of bad debts is; the closer the borrower is to the "lost link" grade, the greater the possibility of losing link; with the extension of the borrowing period, it becomes less likely to keep the original credit quality grade and the transition probability to the next level will gradually increase.

Key words: TEI@I methodology, credit rating and credit scoring, credit quality transfer probability, loss probability prediction, C-K equation

庞素琳, 侯鲜艳. 基于TEI@I方法论的借款人信用品质转移概率计算及失联概率预测应用[J]. 管理评论, 2020, 32(7): 267-279.

Pang Sulin, Hou Xianyan. Calculating Method of Borrower's Credit Quality Transfer Probability and Application in Forecasting Loss-of-Connection Probability Based on TEI@I Methodology[J]. Management Review, 2020, 32(7): 267-279.

参考文献

[1] 汪寿阳,余乐安,黎建强. TEI@I方法论及其在外汇汇率预测中的应用[J]. 管理学报, 2007,4(1):21-27
[2] 余乐安,汪寿阳,黎建强. 外汇汇率与国际原油价格波动预测:TEI@I方法论[M]. 长沙:湖南大学出版社, 2006
[3] 许利枝,汪寿阳. 港口物流预测研究:基于TEI@I方法论[J]. 交通运输系统工程与信息, 2012,12(1):173-179
[4] Wang T., Tian X., Yu M. Stage Division and Pattern Discovery of Complex Patient Care Processes[J]. Journal of Systems Science and Complexity, 2017,5:1136-1159
[5] Carlos., Serrano-Cinca., Begoña., et al. Determinants of Default in P2P Lending[J]. PLOS ONE, 2015,10(10):1-22
[6] 庞素琳,王立. 信用贷款风险中反向CDS协议设计与定价模型[J]. 管理科学学报, 2016,19(6):114-124
[7] Lin M., Prabhala N., Viswanathan S. Judging Borrowers by the Company They Keep:Friendship Networks and Information Asymmetry in Online Peer-to-Peer Lending[J]. Social Science Electronic Publishing, 2013,59(1):17-35
[8] 王性玉,任乐,赵辉,等. 农户信誉特征、还款意愿传递与农户信贷可得——基于信号传递博弈的理论分析和实证检验[J]. 管理评论, 2019,31(5):77-88
[9] 李焰,高弋君,李珍妮,等. 借款人描述性信息对投资人决策的影响——基于P2P网络借贷平台的分析[J]. 经济研究, 2014,49(S1):143-155
[10] 刘巧莉,温浩宇,Hong Qin. P2P网络信贷中投资行为影响因素研究——基于拍拍贷平台交易的证据[J]. 管理评论, 2017,29(6):13-22
[11] 姚铮,胡梦婕,叶敏. 社会网络增进小微企业贷款可得性作用机理研究[J]. 管理世界, 2013,(4):135-149
[12] 赵志冲,迟国泰,潘明道. 基于信用差异度最大的信用等级划分优化方法[J]. 系统工程理论与实践, 2017,37(10):2539-2554
[13] 迟国泰,李鸿禧,潘明道. 基于违约鉴别能力组合赋权的小企业信用评级——基于小型工业企业样本数据的实证分析[J]. 管理科学学报, 2018,21(3):105-126
[14] 庞素琳,巩吉璋. C5.0分类算法及在银行个人信用评级中的应用[J]. 系统工程理论与实践, 2009,29(12):94-104
[15] 潘明道,周颖,迟国泰,等. 基于Fisher判别的小型工业企业债信评级模型及实证[J]. 管理评论, 2018,30(3):15-28
[16] 庞素琳,何毅舟,汪寿阳,等. 基于风险环境的企业多层交叉信用评分模型与应用[J]. 管理科学学报, 2017,20(10):57-69
[17] Bekhet H. A., Eletter S. F. K. Credit Risk Assessment Model for Jordanian Commercial Banks:Neural Scoring Approach[J]. Review of Development Finance, 2014,4(1):20-28
[18] 胡心瀚,叶五一,缪柏其. 上市公司信用风险分析模型中的变量选择[J]. 数理统计与管理, 2012,31(6):1117-1124
[19] 王春峰,万海晖,张维. 基于神经网络技术的商业银行信用风险评估[J]. 系统工程理论与实践, 1999,19(9):24-33
[20] Hájek P. Municipal Credit Rating Modelling by Neural Networks[J]. Decision Support Systems, 2011,51(1):108-118
[21] Malekipirbazari M., Aksakalli V. Risk Assessment in Social Lending Via Random Forests[J]. Expert Systems with Applications, 2015,42(10):4621-4631
[22] Yao X., Crook J., Andreeva G. Support Vector Regression for Loss Given Default Modelling[J]. European Journal of Operational Research, 2015,240(2):528-538
[23] 张卫国,卢媛媛,刘勇军. 基于非均衡模糊近似支持向量机的P2P网贷借款人信用风险评估及应用[J]. 系统工程理论与实践, 2018,38(10):2466-2478