PREDICTION OF LATE INSTALLMENT PAYMENT BASED ON INTEGRATION MODELING
CLUSTER ANALYSIS AVERAGE AND WARD LINKAGE WITH SURVIVAL ANALYSIS (CASE
STUDY: HOUSE OWNERSHIP LOAN BANK X CUSTOMERS)
Abstract
Cluster analysis is an exploratory analysis that is used to group
objects into several clusters, which clusters have different
characteristics. Survival analysis with Extended Cox regression is used
when there is a time-dependent predictor variable so that the
proportional hazard assumption is not met. This study integrates the two
methods. The variables used are Collateral, Character, Capacity,
Condition, and Capital (5C), Credit Collectability, and Credit Payment
Time. The 5C variable has many indicators. The data used is secondary
data obtained from a bank. The purpose of this study was to compare the
Extended Cox Regression model based on the integration of Cluster
analysis on Ward, and Average linkage with Survival analysis using the
Extended Cox Regression method. The results showed that the integrated
cluster model in Ward Linkage-based Extended Cox regression was the best
method with two clusters formed and the smallest mean squared error
value of the model, which was 0.265.