ENHANCING CREDIT SCORING PREDICTION IN ISLAMIC BANKING WITH RANDOM FOREST MACHINE LEARNING MODEL : THE ROLE OF MARITAL STATUS

Abstract

This study explores the application of machine learning techniques, particularly the Random Forest algorithm, to predict default risk in Islamic consumer financing, with a specific focus on marital status as a key demographic factor. Conducted in the context of Islamic banking in Indonesia where ethical compliance and prudent risk assessment are critical the research examines whether incorporating marital status can improve credit risk classification. Utilizing historical financing data from an Islamic bank, the study addresses three central research questions: (1) How accurate is the Random Forest model in predicting default risk when marital status is considered? (2) How effective is the Random Forest algorithm in identifying default risk for Islamic consumer financing based on marital status? (3) What marital status related factors significantly influence the performance of the Random Forest model in this context? The methodology involves standard machine learning procedures, including data preprocessing, categorical feature encoding, and model evaluation using confusion matrices and classification metrics. Feature importance analysis is also conducted to identify influential variables. This research contributes to the emerging synergy between Islamic finance and artificial intelligence, demonstrating how demographic factors such as marital status can enhance Sharia-compliant credit risk assessments in modern Islamic banking systems.