受到新巴塞爾資本協定（New Basel II Accord）將在2006年底實施的影響，除了主管機關及銀行公會已組小組研議因應對策，國內金融機構也逐漸重視此新挑戰，紛紛進行相關的規劃工作，以因應此革命性的變革，因此各大銀行對風險管理等相關課題日益重視。除此之外，近年來國內個人金融消費性貸款在近年來快速的成長，根據行政院金融監督管理委員會統計，民國93年12月已動用的信用卡張數為4418.2萬張，而93年12月20歳以上的人口為1670.9萬人，等於說平均每人持有卡數為2.64張。到了民國94年1月的時候已動用信用卡的張數則已經成長為4451.1萬張，循環信用總餘額為新台幣4636億元，簽帳總金額為1250億元，預借現金總金額為196億元，成長速度驚人。有鑑於此，本研究從個人消費金融著手（以汽車貸款為例），期望利用卡方獨立性檢定、羅吉斯迴歸分析、存活分析和決策樹等統計方法找出顯著影響貸款品質的因子，並建立一個貸款客戶評分標準。

　　研究結果顯示重要的影響因素為：職業、性別、貸款金額、貸款比例、貸款人年齡和汽車價格，將這六個變數留下來之後，以勝算比（odds ratio）的概念建立評分標準，再以四分位差的方法將母體資料分成三等份，命名為A、B和C三等級，其中A為信用評等最佳的客戶、C為信用評等最差的客戶。研究最後並再進一步以決策樹分析了解在C級客戶中，又可以「性別」找出貸款信用品質尚可接受的客戶並予以核貸。

　With New Basel II Accord going to take effect at the end of 2006, many efforts have been made in Taiwan both at the public and private sectors in order to be confronted with the foreseeable challenges. To this end governmental ministry officers, along with the banking industry, have been drafting various strategic plans and proposals to hopefully cope with the revolutionary change. Therefore, the banking industry has become more and more conscious of risk management issues.

　Parallel to this change is the rapid growth of personal consumption credits in recent years, which could cause national credit crisis. According to the statistics published by the Financial Monitoring Committee of the Executive Yuan, the number of credit cards in use had reached 441,820 (in thousands) in December 2004. If this is calculated against the population of people over twenty years of age, the number is 16,709 (in thousands); that is, each adult has approximately an average of 2.64 credit cards. By January 2005, the number of credit cards in use had grown to 445,110 (in thousands) with a circulating credit of 463,600 million dollars, a total signed payment of 125,000 million dollars, and a cash loan of 19,600 million dollars, all with incredible growth rate. To deal with the problem of credit risk, this study tries first to find factors affecting the quality of loan, and then build up criteria for customer credit assessment. Taking car loan as an example, the author explores the nature of risk in personal consumption credits using statistical methods such as chi-square test for independence, multinomial logistic regression for credit prediction, survival data analysis, and decision tree.

　The research result shows that there are six important factors involved in a loan risk. They are: occupation, sex, credit amount, credit ratio, customer age, and car price. With these six factors extracted, the author uses odds ratio to build up the assessment criteria, and then applies quartile deviation to divide the matrix into three equal parts, namely, A, B, and C with A being the most credit customers and C the worst credit customers. In the final stage of the research, the C class customers are analyzed by the regression tree. It is found that “age” is a reliable variable to ensure an acceptable credit quality for crediting.

參考文獻
一. 英文部分
1. Agresti, A. (2002), Categorical Data Analysis -Second Edition, Wiley.
2. Altman, E. I. (1968), ”Financial Ratios, Discriminant Analysis, and the
Prediction of Corporate Bankruptcy,” Journal of Finance, 23, 589-609.
3. AnswerTree 3.0 User’s Guide, SPSS Inc.
4. Jerald, F. L. (2003), Statistical Models and Methods for Lifetime Data, Second
Edition, Wiley.
5. Lee, S. H. and Urrutia, J. L. (1996), ”Analysis and Prediction of Insolvency
in the Property-Liability Insurance Industry：A comparison of Logit and Hazard
Model,” TheJournal of Risk and Insurance, 63, 121-130.
6. Breiman, L., Friedman, J. H., Olshen, R. A. and Stone, C. J. (1984),
Classification and Regression Trees, Wadsworth.
7. Shumway, T. (2001), “Forecasting Bankruptcy More Accurately, A Simple Hazard
Model,” The Journal of Business, Vol.74.1, 101-124.
8. Vandell, K. D., Barnes, W., Hartzell, D., Kraft, D. and Wendt, W. (1993),
“Commercial Mortgage Default：Proportional Hazards Estimation Using Individual
Loan Histories,” Journal of American Real Estimate and Urban Economics
Association, Vol.21.4, 451-480.
9. Zmijewski, M. E.(1984), ”Methodological Issues Related to the Estimation of
Financial Distress Prediction Model,” Journal of Accounting Research, 22,
59-86.
二. 中文部分
1. 王濟川、郭志剛（2003），Logistic迴歸模型－方法及應用，五南圖書。
2. 呂金品（2001），運用決策樹分類方法估計小額放款倒帳率－兼論放款策略，國立中山大
學財務管理研究所碩士論文。
3. 呂美慧（2000），金融機構房貸客戶授信評量模式分析－Logistic迴歸之應用，政治大學
金融研究所碩士論文。
4. 林秀娟、張紹評、張紹勳（2002），SPSS For Windows統計分析、初等統計與高等統計第
四版，松崗電腦圖書資料。
5. 林建州（2001），銀行個人消費信用貸款授信風險評估模式之研究，中山大學財務管理學
系研究所碩士論文。
6. 陳家豪（2003），存活分析方法應用於汽車貸款客戶信用風險管力之研究，成功大學統計
研究所碩士論文。
7. 張文生（2001），銀行建構信用卡信用風險即時預警系統之研究，中原大學企業管理研究
所。
8. 黃俊英（1992），行銷研究－管理與技術，華泰書局。
9. 黃俊英（1991），多變量分析第四版，中國經濟企業研究所。
10. 賴世培、丁庭宇、莫季雍、夏學理（1996），民意調查，國立空中大學。
11. 簡安泰（1977），消費者信用評分制度之研究，政治大學企研所碩士論文。