Model Data Mining untuk Penetapan Plafon Kredit dengan Algoritma C4.5

Frans Mikael Sinaga
Jefri Junifer Pangaribuan
Noor Asegaff, Aulia Rizky Muhammad Hendrik
Chandra, Wenripin
Riche, Riche

Jurnal Sains dan Teknologi
Vol. 4 No. 6 (2025): Desember 2025

Manual credit limit determination in distributor companies is often subjective and inconsistent, increasing the risk of bad debts. This research aims to design an objective data mining model to support customer credit limit decisions at CV. XYZ. The method used is the Decision Tree with the C4.5 algorithm, applied to 66 historical records of customer payment data. Data analysis was performed by calculating Entropy and Information Gain values to build the decision tree, which was then validated using RapidMiner Studio software. The research successfully built a valid and consistent classification model. The "Piutang" (receivables/transaction volume per invoice) attribute was identified as the main determinant (root node), followed by the "Pembayaran" (payment history) attribute as a branch node. This model generates three interpretable decision rules, including the discovery of a risky pattern where high-volume customers with poor payment histories are associated with large credit limits. The proposed model can be implemented as a decision support tool to standardize credit policies, reduce subjectivity, and minimize the company's financial risk.