LOAN FRAUD DETECTION USING DECISION TREE AND RANDOM FOREST MODELS

Banking system vulnerabilities generated opportunities for fraudulent activities, which result in both financial losses and reputational harm for banks and their customers. Financial fraud in the banking sector each year results in significant monetary losses. The continuous problem led financial institutions to close multiple banks, which denied potential borrowers access to loans while producing numerous job losses among banking staff. This study leverages past loan fraud records and employs machine learning to detect fraudulent activities in bank loan applications. The integration of data mining technology improves loan administration through deficiency detection in loan applications, before potential future risks that manual credit officer evaluation might miss. In this work, we utilize two machine learning approaches, i.e., Decision Tree and Random Forest.