Loan Defaulter Risk Prediction

Abstract

This project uses a Random Forest Classifier to predict loan default risk, offering a more accurate 
and flexible alternative to traditional rule-based credit scoring. It analyses borrower specific factors 
including income, credit score, DTI ratio, and employment details to predict risk more accurately. 
The predictive analysis is then used to classify applicants into low, medium or high risks 
categories, helping the loan officers to make more efficient in decision making. A Flask-based web 
application integrates the trained model, allowing loan officers to  input borrower data  and 
receive real-time risk prediction with associated probabilities. Data -preprocessing was performed 
using pandas and numpy.