Comparative Analysis of Machine Learning Models in Credit Risk Evaluation

Abstract

Loan approval is an important process in the financial sector, influencing both lenders and applicants. Traditional methods often rely on manual evaluation, which tend to be slow, inconsistent, and prone to bias. To do away with various such limitations, this paper explores employing machine learning techniques for predicting loan approval outcomes. The research compares three widely used classification algorithms: Random Forest, Logistic Regression, Decision Tree . A structured dataset containing applicant details such as financial status, employment information, and credit history was used. Data preprocessing techniques which include: handling absent values, encoding categorical variables, and feature scaling were applied to ensure data quality. The models had to be trained and examined using performance metrics like precision, recall, F1-score, and ROC-AUC. Among the models, Random Forest demonstrated the most reliable performance given its capacity to capture complex patterns and reduce overfitting. The findings showcase the underlying capacity of machine learning in improving decision-making efficiency and accuracy in loan approval systems.