基于小样本的商业银行信用评级模型研究
发布时间：2019-01-11 18:46
【摘要】：商业银行信用评级(Bank's Credit Rating)是对一家银行当前偿付其金融债务的总体金融能力的评价。对商业银行进行信用评级意义重大：第一,商业银行信用的评级结果是各国金融监管当局进行监测和控制银行业风险维护金融体系安全的根据。第二,商业银行信用评级也是银行进行自身风险管理的基础。第三,金融机构间的业务往来、合作关系的建立需要以商业银行信用评级状况作为基础。第四,工商企业和社会公众可以根据各商业银行的信用评级状况选择与其业务往来的商业银行。 本论文共分五章。第一章分析了论文的选题依据、相关研究进展、研究方法、研究的技术路线和研究内容。第二章基于非线性映射的商业银行信用评价指标体系的构建。第三章关于最优赋权方法的商业银行信用评价模型研究。第四章关于小样本问题的信用评级研究。第五章为结论与展望。论文的主要工作如下： (1)引入非线性映射原理,构建了商业银行信用评级指标体系 根据有、无特定指标两种状态的非线性映射结果的欧氏距离,反映特定指标对评价结果的影响程度,解决了特定指标对评价结果影响程度的可量化问题,为非线性映射删除指标奠定基础。对全部指标做非线性映射,删掉特定指标再做非线性映射。求两个非线性映射结果的欧氏距离,欧氏距离无变化或变化很小表示该指标对评价结果影响小。通过设定复相关系数和相关系数同时超过阈值删除指标,避免了单一标准导致指标的误删,保证了删除指标后信息含量损失少。建立了6个准则层19个指标的商业银行信用风险评价指标体系,用17%的指标反映了91%的原始信息。实证研究表明,本研究的评价结果与穆迪、大公国际的评级结果序关系一致。保持与穆迪、大公国际的评级结果序关系一致有两个原因：一是穆迪、大公国际等权威机构的核心评级指标、评价方法以及赋权方法是不对外披露的。因此人们无法通过其方法对权威机构没有评级的商业银行进行评级。二是保证评价结果与权威机构序关系一致,既保证了评价结果的合理性又解决了可以对所有商业银行进行评级的问题。 (2)利用改进Spearman检验,建立了最优赋权方法的商业银行信用评价模型 根据权威评级机构公布的商业银行信用评级结果,利用改进的Spearman秩相关检验对不同赋权方法得到的评价结果进行检验,选择与权威机构评级结果最接近的赋权方法作为最优赋权方法,解决了现有研究无法解决的最优赋权方法确定的问题。采用了主观赋权的AHP、G1法和客观赋权的离差最大化法、变异系数法、熵值法等五种赋权方法。与权威评级机构穆迪和大公国际的评级结果进行对比检验,解决了现有研究无法解决的最优赋权方法确定的问题。实证研究表明,对于商业银行的信用评价问题,使用熵值法赋权最合适。 (3)建立了基于小样本检验——模拟的商业银行评级模型 通过对评价得分的分布进行检验,找到评价得分的分布规律,为评价得分的数据扩充提供依据,这正是本研究区别于现有研究的不同之处。根据通过分布检验的分布参数模拟生成与评价得分同分布的随机数据,扩充样本数量,使得扩充后的样本数据与原始数据具有相同的分布特征,解决小样本无法划分等级的问题。实证研究表明,中国的商业银行的评价得分既不是正态分布,也不是指数正态分布和对数正态分布,而是服从一种特殊分布,评价得分自然对数的平方服从正态分布的。通过这一分布规律,可以对商业银行的评价得分数据进行模拟扩充达到大样本要求,避免了小样本划分评价等级不准确的问题。
[Abstract]:Bank's Credit Rating is an evaluation of the overall financial capacity of a bank to pay its financial debt. The credit rating of commercial banks is of great significance: first, the credit rating of commercial banks is the basis for the supervision and control of the banking risk and the safety of the financial system. Secondly, the credit rating of commercial banks is the foundation of the bank's own risk management. Third, the establishment of the relationship between the financial institutions and the establishment of the cooperative relationship needs to be based on the credit rating of commercial banks. The fourth, the business enterprise and the public can choose the commercial bank of the commercial bank according to the credit rating situation of each commercial bank. This paper is a total of five The first chapter analyses the basis of the selection of the thesis, the research progress, the research methods, the technical route and the research of the research. The second chapter is based on the structure of the index system of the credit evaluation of the commercial bank based on the non-linear mapping The Research on the Credit Evaluation Model of Commercial Banks in Chapter 3 on the Method of Optimal Empowerment Research. Chapter IV Research on the Credit Rating of the Small-sample Problem The fifth chapter is the conclusion and exhibition. The main work of the paper is as follows: (1) The principle of non-linear mapping is introduced, and the credit rating of commercial banks is constructed. The index system is based on the Euclidean distance of the non-linear mapping results of the two states with no specific indexes, and reflects the result of the specific index to the evaluation. The influence degree of the specific index on the evaluation result is solved, and the non-linear mapping deletion finger lay the foundation of the bid. Make a non-linear mapping to all the indexes, and delete the specific indexes to do the same. Non-linear mapping. The Euclidean distance of two non-linear mapping results is obtained. There is no change or change in the Euclidean distance, which indicates the evaluation of the index. The result is small. By setting the complex correlation coefficient and the correlation coefficient at the same time to exceed the threshold value deletion index, the error deletion of the index caused by the single standard is avoided, and the information after the deletion indicator is guaranteed. The content loss is little. The index system of the credit risk evaluation of the commercial banks with the 19 indexes of the six criteria is established, and the index of 17% reflects 91%. The empirical study shows that the results of this study are similar to that of Moody's and the Great Public. The order relationship is consistent. There are two reasons for keeping the order relationship with Moody's and Big Public International: one is the core rating index of the authorities such as Moody's and Big Public International, and the evaluation method and the weighting method are not A commercial silver that cannot be rated by an authority by way of its method. the second is to ensure that the evaluation result is consistent with the authority order relationship, so that the rationality of the evaluation result is ensured, and all commercial banks can be solved The problem of rating. (2) With the improved Spearman test, the business of the optimal weighting method is established. The bank credit evaluation model is based on the results of the commercial bank credit rating published by the authority rating agency, and the improved Spearman rank correlation test is used to obtain the different weighting methods. The result of the evaluation is tested, and the method of weighting which is closest to the rating result of the authority is selected as the optimal weighting method, and the problem that the existing research can not be solved is solved. The problem of the determination of the method of weight is the method for maximizing the difference of the AHP, the G1 method and the objective weight, the coefficient of variation, the entropy, the entropy of the objective weight, the coefficient of variation, the entropy, The five weighting methods, such as the value method, are compared with the rating results of the authoritative rating agency Moody's and the Big Public International to solve the problem that the existing research can not be solved. The empirical study shows that the problem of credit evaluation for commercial banks The method of entropy value is used to empower the most suitable. (3) a small sample test is established. The model of the credit rating of the commercial bank is to test the distribution of the evaluation score, find the distribution law of the evaluation score, and provide the basis for the data expansion of the evaluation score, which is the present research. according to the distribution parameter of the distribution test, the random data which is in the same distribution as the evaluation score is simulated, the number of the samples is expanded, so that the expanded sample data and the original data have the same distribution characteristic, the solution The empirical study shows that the evaluation score of commercial banks in China is neither a normal distribution nor an exponential normal distribution nor a log-normal distribution, but is a special distribution and the evaluation score is self-determined. By this distribution rule, the evaluation score data of commercial banks can be simulated and expanded to meet the requirement of large samples, and the small samples are avoided.
【学位授予单位】：大连理工大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：F831.2
本文编号：2407345
[Abstract]:Bank's Credit Rating is an evaluation of the overall financial capacity of a bank to pay its financial debt. The credit rating of commercial banks is of great significance: first, the credit rating of commercial banks is the basis for the supervision and control of the banking risk and the safety of the financial system. Secondly, the credit rating of commercial banks is the foundation of the bank's own risk management. Third, the establishment of the relationship between the financial institutions and the establishment of the cooperative relationship needs to be based on the credit rating of commercial banks. The fourth, the business enterprise and the public can choose the commercial bank of the commercial bank according to the credit rating situation of each commercial bank. This paper is a total of five The first chapter analyses the basis of the selection of the thesis, the research progress, the research methods, the technical route and the research of the research. The second chapter is based on the structure of the index system of the credit evaluation of the commercial bank based on the non-linear mapping The Research on the Credit Evaluation Model of Commercial Banks in Chapter 3 on the Method of Optimal Empowerment Research. Chapter IV Research on the Credit Rating of the Small-sample Problem The fifth chapter is the conclusion and exhibition. The main work of the paper is as follows: (1) The principle of non-linear mapping is introduced, and the credit rating of commercial banks is constructed. The index system is based on the Euclidean distance of the non-linear mapping results of the two states with no specific indexes, and reflects the result of the specific index to the evaluation. The influence degree of the specific index on the evaluation result is solved, and the non-linear mapping deletion finger lay the foundation of the bid. Make a non-linear mapping to all the indexes, and delete the specific indexes to do the same. Non-linear mapping. The Euclidean distance of two non-linear mapping results is obtained. There is no change or change in the Euclidean distance, which indicates the evaluation of the index. The result is small. By setting the complex correlation coefficient and the correlation coefficient at the same time to exceed the threshold value deletion index, the error deletion of the index caused by the single standard is avoided, and the information after the deletion indicator is guaranteed. The content loss is little. The index system of the credit risk evaluation of the commercial banks with the 19 indexes of the six criteria is established, and the index of 17% reflects 91%. The empirical study shows that the results of this study are similar to that of Moody's and the Great Public. The order relationship is consistent. There are two reasons for keeping the order relationship with Moody's and Big Public International: one is the core rating index of the authorities such as Moody's and Big Public International, and the evaluation method and the weighting method are not A commercial silver that cannot be rated by an authority by way of its method. the second is to ensure that the evaluation result is consistent with the authority order relationship, so that the rationality of the evaluation result is ensured, and all commercial banks can be solved The problem of rating. (2) With the improved Spearman test, the business of the optimal weighting method is established. The bank credit evaluation model is based on the results of the commercial bank credit rating published by the authority rating agency, and the improved Spearman rank correlation test is used to obtain the different weighting methods. The result of the evaluation is tested, and the method of weighting which is closest to the rating result of the authority is selected as the optimal weighting method, and the problem that the existing research can not be solved is solved. The problem of the determination of the method of weight is the method for maximizing the difference of the AHP, the G1 method and the objective weight, the coefficient of variation, the entropy, the entropy of the objective weight, the coefficient of variation, the entropy, The five weighting methods, such as the value method, are compared with the rating results of the authoritative rating agency Moody's and the Big Public International to solve the problem that the existing research can not be solved. The empirical study shows that the problem of credit evaluation for commercial banks The method of entropy value is used to empower the most suitable. (3) a small sample test is established. The model of the credit rating of the commercial bank is to test the distribution of the evaluation score, find the distribution law of the evaluation score, and provide the basis for the data expansion of the evaluation score, which is the present research. according to the distribution parameter of the distribution test, the random data which is in the same distribution as the evaluation score is simulated, the number of the samples is expanded, so that the expanded sample data and the original data have the same distribution characteristic, the solution The empirical study shows that the evaluation score of commercial banks in China is neither a normal distribution nor an exponential normal distribution nor a log-normal distribution, but is a special distribution and the evaluation score is self-determined. By this distribution rule, the evaluation score data of commercial banks can be simulated and expanded to meet the requirement of large samples, and the small samples are avoided.
【学位授予单位】：大连理工大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：F831.2
【引证文献】
相关硕士学位论文 前1条
1 李翘;平台型电子商务诚信评估体系研究[D];厦门大学;2014年
，本文编号：2407345
本文链接：https://www.wllwen.com/guanlilunwen/huobilw/2407345.html
