Analysis Of Financing Risk Optimization At XYZ Multifinance Company Through A Credit Score Clustering Approach

Ranti Ramayanti Haris (a:1:{s:5:"en_US"s:21:"Universitas Gunadarma"})
Astie Darmayantie (Universitas Gunadarma)

International Journal of Health, Engineering and Technology
Vol. 5 No. 2 (2026): Vol 5. No. 2 JULY 2026

This research a credit scoring model to optimize financing risk in Multifinance Company XYZ, which has a Non-Performing Financing (NPF) rate of 2.8%, exceeding the company’s target of 2%. The development is carried out using Oracle Development Suite with the CRISP-DM methodology, covering the stages of business understanding, data understanding, data preparation, modeling, evaluation, and deployment. The dataset consists of 250 customer records from the period 2015–2025, with 101 initial variables that are subsequently reduced to 14 predictive variables based on correlation analysis and credit scoring domain knowledge. A K-Nearest Neighbor (KNN) classification model with K = 3 is developed using Oracle Data Miner to classify borrower risk into three categories: LOW, MEDIUM, and HIGH. Testing results on 50 testing data show that the model achieves an accuracy of 98%, precision of 97.5%, and recall of 100% for the HIGH-risk class. A 5-fold cross-validation confirms the model’s consistency with a mean accuracy of 97.6% (±1.1%). Implementation using PL/SQL and Oracle Forms produces a real-time scoring system with a processing time of less than 5 seconds per application, compared to 5–7 days in the previous manual proces.