臺灣博碩士論文加值系統

根據財政部中央銀行統計，台灣地區個人消費性放款總額已於民國八十五年年底超過法人企業放款，同時個人消費性放款成長率也大於法人企業放款，使得兩者之間的差距也逐年增加中；再加上台灣地區將加入WTO、經濟自由化與國際化、個人投資理財觀念的盛行，個人放款會巨幅的成長是可以預期的結果，也將成為各銀行未來主要的經濟命脈之一。
在歐美各國，許多較具規模的金融機構皆有屬於自己各項產品的信用評分模型及相關系統與研究部門，但國內銀行所走的腳步較為緩慢，雖有信用評分制度的實行，但多為參照它行加以修改而得之，缺乏科學依據；Shanker (1996)的研究顯示，資料標準化可以改善區別率，本研究希望能利用分析層級程序法將專家的對各項變數權重的看法加入類神經網路中，增加信用評分模型之正確性。資料來源為國內某上市銀行之88-89無擔保信用貸款，採問券方式取得分析層級程序法各項權重，做為類神經網路資料標準化的依據，另一方面以線性方式將資料標準化，訓練出另一倒傳遞類神經網路，再發展羅吉斯迴歸法之信用評分模型，最後將三者以測試樣本做正確率比較，
結論發現，結合分析層級程序法之類神經網路有較高的正確率，發展類神經網路信用評分模型時，我們可以使用分析層級法來做為資料標準化的依據，相信會有不錯的效果。

According to a 1996 national survey conducted by the R.O.C. Central Bank in Taiwan, both the amount and the increasing rate of consumer loan are greater than those of corporate loan. Further, joining the World Trade Organization (WTO), personalized financial management, and international business on Internet make Taiwan’s consumer loan grow up hugely and then consumer loan owns a large proportion of each banks’s total assets in the neare future. In this situation, a successful credit scoring model can be used for each bank to classify applicants for credit into “good” and “bad” risk classes, as well as to survive in a competitive banking market.
In a survey of current techniques for solving a credit scoring model in business, the back propagation neural network and logistics regression models dominated other methods (Velido et al., 1999; West, 2000). On the oter hand, data standardization and the mixture of experts could improve the classification rate and accuracy respectively (Shanker, 1997; West, 2000). Therefore, this study integrates the Analytic Hierarchy Process (AHP) and the back-propagation neural network model sequentially. First, experts’ views and wetghrs related to several consumers’ properties are consistently obtained from the AHP process and then the results are used to train the neural network for more accurate credit scoring model. The data sets are from a bank which already runs over twenty years in Taiwan. Experimental results show that the proposed new approach is found to be the most accurate of the three methods descreibed in this study. Finally, some general guidelines and future research directions ate presented.

表 次vi
圖 次ix
第一章 緒論1
第一節 研究動機1
第二節 研究目的4
第三節 研究範圍5
第四節 預定論文結構5
第二章 文獻探討8
第一節 消費者小額信用貸款8
第二節 國內外相關研究13
一、國外部份13
二、國內部份18
第三章 研究方法25
第一節 研究架構25
第二節 理論基礎與使用工具40
一、 羅吉斯迴歸模式(Logistic Regression )40
二、 類神經網路44
三、分析層級程序法48
第三小節 使用工具55
第四章 實證分析56
第一節 羅吉斯迴歸信用評分模型56
第二節 倒傳遞類神經網路模型60
第三節 AHP與類神經網路混合模型61
一、AHP權重之的決定61
二、混合模型之建立69
第五章 結論與建議70
第一節 結論70
第二節　研究限制與建議74
一、研究限制74
二、研究建議75
參考文獻76
附表一、AHP訪談表81

1.P. Henry Mueller著，葉英俊編譯，授信風險管理，台北：財團法人金融人員研究訓練中心，1994，頁31-34
2.中央銀行經濟研究處編，”本國一般銀行放款餘額”，金融統計月報，2000年10月，頁85-86
3.李俊毅，應用灰色預測理論與類神經網路於企業財務危機 預警模式之研究，義守大學管理科學研究所，1999
4.呂美慧，銀行授信評等模式---Logistic Regression 之應用，國立政治大學金融學系碩士論文，2000
5.周文賢著，多變量統計分析-SAS/STAT 使用方法，台北：智勝書局，2001，頁 5-32
6.郭敏華著，債信評等，台北：智勝出版社，2000，頁193-254
7.郭怡萍，銀行業信用評等模式之建構，國立台灣大學商學研究所，1998
8.財團法人金融人員研究訓練中心編，消費者小額信用貸款之銀行授信實務概要，台北：財團法人金融人員研究訓練中心，1999，頁185-190
10.張永豪，銀行信用評等─新銀行之實證分析，朝陽大學財務金融系碩士論文，1999
11.黃小玉，銀行放款信用評估模式之研究--最佳模式之選擇，淡江大學管理科學研究所碩士論文，1989
12.黃登源著，應用迴歸分析，台北：華泰文化事業公司，1998，頁294-320
13.葉怡成編著，類神經網路模式應用與實作，台北：儒林圖書，1993，頁4-1~4-44
14.陳文賢著，管理科學--作業研究與數量方法，台北：三民書局，1987，頁71-101
15.簡安泰，消費者信用評分制度之研究，政大企研所碩士論文，1977
16.謝維正，台灣股票上市公司信用評等區別模式之探討，國立交通大學管理科學研究所碩士論文，1987
17.鄭瑞楠，財務分析在銀行授信決策上的應用之研究，國立東華大學企業管理學系碩士論文，1998
18.龐欽元，農業行庫授信風險評估模式之研究，國立中興大學農業經濟學系碩士論文，1999
19.Altman EI., “Corporate distress diagnosis : comparisions using linear descriminant analysisi and neural networks(the Italian experience),” Journal of Banking and Finance, 18,1994, 524-529.
20.Abdel-Khalik AR, El-Sheshai KM, “Information choice and utilization in an experiment on default prediction,” Journal of Accounting Research, Autumn 1980, 325-342.
21.Carter. C. and Catlett. J., “Assessing credit card applications using machine learning, ” IEEE Expert, fall , 1987, 71-79.
22.Coats PK,Fant LF., “Recognizing financial distress patterns using a neural network tool,” Financial Management, Autumn 1993, 142-155.
23.Coakley J. R. and Brown C. E., “Artificial Neural Networks in Accounting and Finance: Modeling Issues,” International Journal Intelligent System in Accounting Finance Management, 9, 2000 , 119-114.
24.Coffman, J. Y., “The proper role of tree analysis in forecasting the risk behavior of borrowers,” MDS Reports 3,4,7 and 9 Management Decision Systems, Atlanta, 1986.
25.David West., “Neural network credit scoring models,” Computer & Operations Research, 27 , 2000 , 1131-1152
26.Davis, D. B., “Artificial intelligence goes to work,” High Technol, Apr, 1987, 16-17.
27.Davis RH, Edelman DB, Gammerman AJ., “Machine learning algorithms for credit-card application,” IMA Journal of Mathematics Applied in Business and Industry, 4, 1992, 43-51
28.Desai VS, Crook JN, Overstreet GA, “A comparison of neural networks and linear scoring models in the credit union environment,” European Journal of Operational Research, 95, 1996, 24-37
29.Durand, D., “Risk Elements in Consumer Instalment Financing,” New York: National Bureau of Economic Research, 1941.
30.Fitzpatrick, D. B., “An analysis of bank credit card profit,” J. BankRes, 7, 1976,199-205.
31.Grablowsky, B. J. and Talley, W. K., “Probit and discriminant functions for classifying credit applicants: a comparison,” Journal Economic Business, 33, 1981, 254-261.
32.Hand D.J. and Henley W.E., “Statistical Classification Methods in Consumer Credit Scoring: A Review,” J.R. Statist. Soc. A., 160. Part3 , 1997, 523-541.
33.Henley, W. E.,“Statistical aspects of credit scoring,” PhD Thesis, The Open University, Milton Keynes, 1995.
34.Jensen H.L., Robert R. Trippi, Efraim Turban , Neural networks in finance and investing : using artificial intelligence to improve real-world performance, eChicago, Ill. : Probus Pub., 1993.
35.Jensen H.L., “Using neural networks for credit scoring,” Managerial Finance, 18, 1992, 15-26.
36.Kolesat, P. and Showers, J. L., “A robust credit screening model using categorical data,” Management Science, 31, 1985, 123-133.
37.Leonard, K. J., “Detecting credit card fraud using expert systems,” Comput. Indstrl Engng,25, 1993a, 103-106.
38.Leonard, K. J., “A fraud-alert model for credit cards during the authorization process,” IMA J. Math Appl, Bus Indstry,5, 1993b, 57-62.
39.Lucas, A., “Updating scoredcards:removing the mystique. In Credit Scoring and Credit Control(eds L. C. Thomas, J. N. Crook and D. B. Edelman),” Oxford: Clarendon, 1992,180-197
40.George Curnow, Gary Kochman, Steven Meester, Debashish Sarkar, Keith Wilton, “Automating Credit and Collections Decisions at AT&T Capital Corporation,” Management Science, January-February 1997, 29-52.
41.Makowski, P., “Credit scoring branches out,” Credit World, 75, 1985,30-37.
42.Oliver, R. M., “The economic value of score-splitting accept-reject policies,” IMA J. Math. Appl. Bs. Indstry, 4, 1992, 35-41.
43.Orgler,Y.E., “A credit scoring model for commercial loans,” J. Money Credit Bank, Nov, 1970, 435-445.
44.Piramuthu S, “Financial credit risk evaluation with neural and neurofuzzy systems,” European Journal of Operational Research, 112, 1999, 310-321.
45.Quinlan JR., “Simplifying decision trees,” International Journal of Man-Machine Studies, 27 , 1987，221-234.
46.Saaty T.L.,The Analytic Hierarchy Process, New York：McGraw-Hill，1980.
47.Shanker M, Hu MY, Hung MS, “Effect of data standardization on neural network training,” Omega, international Journal of Management Science, 24,No. 4, 1996 , 385-397.
48.Showers, J. L. and Chakrin, L. M., “Reducing uncollectable revenue from residential telephone customers,” Interfaces, 11, 1981, 21-31.
49.Tam KY., Kiang MY. , “Managerial applications of neural network : the case of bank failure predictions,” Management Science, 1992, 38, 926-947.
50.Zocco,D.P., “A framework for expert systems in bank loan management,” J. Commrcl Bank Lend, 67, 1985, 47-54.