| 摘要: | 台中市地區L銀行，中華民國95年至104年的房貸款客戶授信評估是本研究的課題，隨機抽樣為323案例，其中295戶為正常房屋貸款客戶，28例為異常情況。17項變數經文獻綜述，實際經驗選擇（年齡，性別，學歷，婚姻狀況，職稱，工作，服務年資，還款方式，擔保人，收入，擔保品所在地區，擔保品類型，貸款目的，是否有兩套以上的住房貸款，無論是擔保債務，信用卡還是現金卡有循環餘額，由其他金融機構檢查）選擇羅吉斯迴歸模型，並設置為經驗模型(LR1)，另將該17個風險變數進行羅吉斯迴歸選取，篩選出之顯著變數有：學歷，職稱，服務年資，償還方式，擔保品地點，保證債務，信用卡或現金卡雙卡循環動用，並設置為實證模型(LR2)。 LR1和LR2模型由Logistic迴歸分析，其分類準確率分別為94.43％和93.19％，LR1模型預測正常抵押貸款客戶的正確率約為99.66％，正常抵押貸款客戶的正確預測率約為39.29％ LR1模型顯示出更好的預測能力，選擇LR1是最好的信用風險評估模型。
The housing loan customers in the period time of the Republic of China 95 years to 104 years of a selected L bank in Taichung city were the subjects for this study, by random sampling for the study 323 sample, of which 295 were normal housing loan customers and 28 for the abnormal ones. The 17 variable through literature review, practical experience selected (age, gender, education level, marital status, job title, job, years of service, repayment, the guarantor, income, property located in the region, property type, loan purpose, whether two or more housing loans, whether guaranteed debt, credit card or cash card have cycle balance, checked by other Financial institution) are entered into the logistic regression model, and set LR1, then install 17 variables in the logistic regression to filter a significant variable, and set LR2(significant variable are: education level , job title, years of service, repayment, property located in the region, guaranteed debt, debt, credit card or cash card have cycle. and set LR2. ) LR1 and LR2 model are processed by The logistic regression, which totle classification accuracy rates were 94.43% and 93.19%, and LR1 model to predict the correct rate of normal mortgage customers about 99.66%, correct prediction rate of abnormal mortgage customers about 39.29%, LR1 nodel shows better prediction capability, select the LR1 is the best credit risk assessment model. |
