Every year, the number of cooperatives in the province of East Nusa Tenggara continues to grow.
Cooperatives are present with the aim of helping the
community on the financial side. The cooperative offers the
principle of saving and providing low-interest loans to its
members. But there are times when lending is subjective. This
condition is a major factor in the occurrence of errors in
providing credit that leads to congestion (non-performing
loans). This study focuses on the comparison of five
classification methods using historical loan application data for
a Multipurpose Cooperative in East Nusa Tenggara. The 5
methods are Naïve Bayes, K-Nearest Neighbor (KNN), Support
Vector Machine (SVM) Random Forest, and C4.5. In the test
results, it turns out that the C4.5 Method has better accuracy
and a smaller error rate.