Credit Scoring Analysis using LASSO Logistic Regression and Support Vector Machine (SVM) Sari Putri Dina, Aidi Muhammad Nur, Sartono Bagus Department of Statistics, Bogor Agricultural University, INDONESIA Online published on 31 October, 2017. Abstract Credit scoring analysis using logistic regression is less effective, as it requires hypothesis and assumption testings that must be fulfilled. Therefore, LASSO logistic regression analysis can be used to overcome this problem because it does not require hypothesis and assumption testings. In addition, Support Vector Machine (SVM) method is also a classification method that has good classification capability as well as can ignore all assumptions such as logistic regression. This study aims to study the factors that affect the smoothness of debtors in paying motorcycle credits and compare the goodness of both methods used. Explanatory variables that affect the smoothness of credit payments are the phone ownership, down payment, loan term, occupation, age, marital status, gender, education level, number of dependents, motorcycle type, interaction between motorcycle type with phone ownership and down payment, interaction between phone ownership with occupation, loan terms, and phone number ownership, interaction between downpayment with loan terms, installments, gender and educational level, and interaction between occupation with tyype of income. Application of LASSO logistic regression and SVM in this case have mostly the same classification accuracy. However, the results of the classification performance using SVM method is relatively stable compared with LASSO logistic regression Top Keywords logistic regression, LASSO, classification, support vector machine. Top |