本文關(guān)鍵詞：基于信用差異度最大的信用等級劃分優(yōu)化方法　出處：《系統(tǒng)工程理論與實踐》2017年10期 　論文類型：期刊論文

  更多相關(guān)文章： 信用評級 信用等級劃分 最優(yōu)劃分 違約金字塔 信用差異度

【摘要】：信用評級對當(dāng)代社會有極其重要的影響,若信用等級劃分不合理,必將誤導(dǎo)債權(quán)人和社會公眾.信用評級結(jié)果的變動直接反映經(jīng)濟狀態(tài)的變化,2011年標(biāo)準(zhǔn)普爾把美國的主權(quán)信用評級從AAA級降為AA+,引起全球金融市場的動蕩.信用評級的本質(zhì)是合理區(qū)分客戶的信用狀況,揭示不同等級客戶的信用風(fēng)險水平.國際上比較流行的標(biāo)普、穆迪的信用評級針對中國客戶的評級結(jié)果往往存在信用等級很高、違約損失率反而不低的不合理現(xiàn)象.本研究以信用差異度和違約金字塔為標(biāo)準(zhǔn),構(gòu)建非線性規(guī)劃模型劃分信用等級,并以中國小企業(yè)貸款數(shù)據(jù)為樣本進行實證研究·本研究的創(chuàng)新與特色一是根據(jù)第k個信用等級中最后一個樣本的信用評分P_(mk)~k與第k+1個信用等級中第一個樣本的信用評分P_1~(k+1)確定相鄰兩個等級的信用評分差值,以所有信用等級的評分差值之和∑(P_(mk)~k-P_1~(k+1))最大為目標(biāo)函數(shù),確保最大程度的保證信用評分差異大的客戶劃分為不同信用等級.避免了把信用狀況差異較大的客戶劃分成同一個信用等級的不合理現(xiàn)象.二是以信用等級由高到低的違約損失率嚴(yán)格遞增為約束條件建立信用等級劃分模型,保證信用等級劃分結(jié)果滿足信用等級越高、違約損失率越低的違約金字塔標(biāo)準(zhǔn),避免出現(xiàn)信用等級很高、違約損失率反而不低的不合理現(xiàn)象.三是1814筆工業(yè)小企業(yè)貸款數(shù)據(jù)的實證研究表明,本研究的信用等級劃分方法不僅滿足信用等級越高、違約損失率越低的違約金字塔標(biāo)準(zhǔn),還能保證信用狀況差異大的客戶劃分為不同信用等級.
[Abstract]:Credit rating has an extremely important impact on contemporary society. If the credit rating is unreasonable, it will mislead creditors and the public. The change of credit rating results directly reflects the change of economic state. In 2011, Standard & Poor's downgraded the United States' sovereign credit rating from AAA to AA, which caused turmoil in the global financial market. The essence of the credit rating is to reasonably distinguish the credit status of customers. To reveal the credit risk level of customers of different grades. The credit rating of S & P and Moody's, which is popular in the world, often has a high credit rating for Chinese customers. Based on the standard of credit difference and default pyramid, the nonlinear programming model is constructed to classify credit grade. The innovation and characteristics of this study are based on the credit score of the last sample of the k credit grade. The credit score of the first sample of the first sample of the k1 credit rating, PHS 1 / k 1) determines the credit score difference between the two adjacent grades. Take as the objective function the maximum of the sum of the rating differences of all credit grades 鈭，

本文編號：1358622