Perbandingan Metode Diskretisasi Dalam Model Regresi Logistik (Studi Kasus : Pembentukan Model Penskoran Kredit Bank X)

Abstract

The number of KPR customers of bank X decreased since February 2012 until August 2012. Logistic regression which is used in credit scoring model is not only able to identify the significance factors but also to know the scoring value in each category of explanatory variable. The data used in credit scoring model must be categorical data. Continuous data needs to be discretized in order to get categorical data. Discretization process can use chimerge method (model I) and equal with interval method (model II). Based on logistic regression, the factors that affect the respon are sex, employment status, education, occupation, age, and income. Correct classification table, shows that model I is better to classify than model II, therefore the scorecard is made based on model I.