<![CDATA[This study aims to develop a credit risk prediction model for customers at PT Toyota Astra Financial Services using data mining algorithms, specifically Random Forest and XGBoost. In response to the challenge of non-performing loans (NPL), machine learning-based predictive models offer an effective solution to identify potential risks early. The research utilizes historical customer data encompassing demographic information, employment status, and loan history. After data preprocessing, the models were evaluated using accuracy, precision, recall, F1-score, and ROC-AUC metrics. The results indicate that XGBoost outperformed other models with an accuracy of 91.67% and an F1-score of 0.89 for the positive class. These findings demonstrate that applying machine learning algorithms can significantly enhance credit selection efficiency and reduce potential losses from defaulted loans.Keywords: Credit Risk; Machine learning; Random Forest; XGBoost, Data mining. AbstrakPenelitian ini bertujuan untuk membangun model prediksi risiko kredit nasabah pada PT Toyota Astra Financial Services dengan memanfaatkan algoritma data mining, khususnya Random Forest dan XGBoost. Dalam menghadapi tantangan kredit macet, model prediktif berbasis machine learning dapat memberikan solusi yang efektif untuk mengidentifikasi potensi risiko sejak dini. Penelitian ini menggunakan data historis nasabah yang mencakup informasi demografi, status pekerjaan, dan riwayat pinjaman. Setelah melalui tahap pra-pemrosesan data, model dievaluasi menggunakan metrik akurasi, presisi, recall, F1-score, dan ROC-AUC. Hasil menunjukkan bahwa XGBoost memiliki performa terbaik dengan akurasi sebesar 91,67% dan F1-score 0,89 pada kelas positif. Temuan ini menunjukkan bahwa penerapan algoritma machine learning dapat meningkatkan efisiensi seleksi kredit dan mengurangi potensi kerugian akibat kredit bermasalah.Kata kunci: Risiko Kredit; Machine learning; Random Forest; XGBoost; Data mining]]>
