Research: "Predicting Loan Approval with Random Forest Algorithm"
Abstract: In this research, I developed a machine learning model to automate the credit eligibility evaluation process. By leveraging the Random Forest algorithm, this study aims to provide fast and accurate predictions to determine whether a customer's loan application should be approved or rejected based on historical data.
Financial institutions face significant risks regarding non-performing loans. Manual evaluation processes are often time-consuming and prone to subjectivity. Therefore, I focus on utilizing predictive analytics to minimize financial risks and enhance banking operational efficiency in processing massive loan applications.
This research follows a systematic Data Science workflow, including:
I designed a measurable research pipeline, ranging from data cleaning to model validation. The primary focus at this stage is feature selection to ensure that the variables most influential to loan status (such as Credit History) receive the appropriate weighting within the model.
Based on the tests conducted, the Random Forest model demonstrated highly effective performance. The model is capable of identifying high-risk customer patterns with a competitive accuracy rate compared to traditional methods, providing a data-driven foundation for risk managers to make informed decisions.
The implementation of algorithms across four model scenarios—Decision Tree, Random Forest, XGBoost, and Random Forest using different datasets—has proven to provide strategic value for financial institutions. I conclude that this automation not only accelerates administrative processes but also significantly reduces potential losses resulting from human error in credit risk analysis.