Cloud-Native Credit Risk Analytics Using Sas Viya and Distributed Data Platforms

Abstract

The rapid growth of digital banking and financial services has increased the need for advanced credit risk management solutions capable of processing large volumes of data efficiently and accurately. This study proposes a cloud-native credit risk analytics framework using SAS Viya and distributed data platforms to enhance credit risk assessment and borrower default prediction. A hypothetical quantitative research approach was adopted, utilizing large-scale financial datasets containing customer demographic information, transaction records, credit history, and repayment behavior. Data preprocessing and feature engineering techniques were employed to improve data quality and predictive performance. Machine learning models, including Logistic Regression, Decision Tree, Random Forest, and Gradient Boosting, were developed and evaluated within the SAS Viya environment. The results indicated that Gradient Boosting achieved the highest predictive accuracy of 93.5% and an AUC value of 0.96, demonstrating excellent capability in distinguishing high-risk and low-risk borrowers. Furthermore, the cloud-native architecture improved scalability, reduced processing time, and enabled efficient distributed data processing. Key predictors of credit risk included credit score, repayment history, debt-to-income ratio, and credit utilization ratio.