摘要
　　隨著科技的進步，大數據時代的來臨，資料分析成為了目前金融業的顯學。但隨著個資保護意識抬頭，能被用來作數據分析的資料逐漸變得難以取得。本研究在探討如何利用授信的資料去做信用評等。本研究用來做信用評等的方式為建構一個信用評分模型，而這個模型是透過受限資料去建構的。模型的建構包含了最後的模型檢驗，用以檢測此方法所建構出的信用評分模型是否為一個有效模型。

關鍵字：信用風險、信用評分卡、羅吉斯迴歸

Abstrac
With the advancement of science and technology, data analysis has become the predominant learning of the current financial industry. However, with the rising awareness of conservation, data that can be used for analysis has gradually become difficult to obtain. This study explores how credit information can be used to make credit ratings. The method used for credit rating in this study is to build a credit scoring model which is constructed through restricted data. The construction of the model includes the final model test to test whether the credit scoring model constructed by this method is a valid model or not.

Keywords: credit risk, credit scorecard, logistic regression

1. 翁慈宗 (2009)，「資料探勘的發展與挑戰」，《科學發展期刊》，442, 33-39。
2. 單良 (2010)，「信用評等模型的十二堂課」，《財團法人台灣金融研訓院出版》。
3. 呂美慧 (2000) ，「銀行授信評等模式- Logistic Regression之應用」，政治大學金融研究所碩士論文。
4. Hubert W. Lilliefors (1967),“On the Kolmogorov-Smirnov Test for Normality with Mean and Variance Unknown,”Journal of the American Statistical Association，318, 399-402.
5. ŘEZÁČ M. and F. ŘEZÁČ (2011)，“How to Measure the Quality of Credit Scoring Models,”Finance a Uver, 61, 486-507.