任何信貸業務都有一定的風險，而信用卡業務又具有無擔保抵押的特性，所以信用卡是一種高風險的金融商品，因此，識別信用卡違約行為的影響因素，並據此對信用卡持有者的違約風險進行衡量和預測，將有助於銀行防範和化解信用卡風險。然而信用卡各項基本資料、銀行通路及刷卡行為等變數與違約的發生，有相當程度的關係。 本研究是以某銀行的信用卡流通卡客戶數為母體樣本，再加上違約戶與正常戶1:1及1:3比例的抽樣樣本數，總共區分為三種樣本數；再搭配「基本資料」、「銀行通路」及「刷卡習慣」三類型自變數，共有七種不同的自變數組合，所進行各式組合之羅吉斯迴歸模型分析，以探究其最適的羅吉斯回歸分析模型，並確認那些變數是影響持卡人違約的重要因素。 ;Any credit business has certain risks, and the credit card business has the characteristics of unsecured mortgage. Therefore, credit card is a high-risk financial product. Therefore, it identifies the influencing factors of credit card default behavior and accordingly defaults on credit card holders. Risk measurement and forecasting will help banks prevent and resolve credit card risks. However, the basic data of credit card, bank access and credit card behavior have a considerable degree of relationship with the occurrence of default. This study is based on the number of credit card circulation card customers of a bank as the parent sample, plus the number of sampled samples of the ratio of 1:1 and 1:3 between the defaulting household and the normal household. The total number of samples is divided into three types of samples; Three types of independent variables, "banking channel" and "swipe habits", there are seven different combinations of independent variables, and the combination of various types of Logis regression model analysis to explore its optimal Logistic regression analysis model, and Confirming those variables is an important factor affecting cardholder defaults.
