Lauri Soome

Data & Engineering

View or Download PDF
← Back to Home

AI Loan Approval Predictor

Overview

This Logistic Regression model achieves 91.14% accuracy with a ROC-AUC score of 0.90, and was trained on over 58,000 loan applications.

Built using Python scikit-learn for model development, FastAPI for the backend service.

Interactive Demo

Try the loan predictor with different applicant profiles. The model analyzes multiple factors to provide real-time approval predictions.

Personal Information

Loan Information

Model Performance

91.14%
Accuracy
0.90
ROC-AUC Score
58K+
Training Samples

The model demonstrates strong predictive performance with high accuracy and ROC-AUC score.

Technical Implementation

Data Science Pipeline

  • Feature engineering with one-hot encoding for categorical variables
  • StandardScaler normalization for numeric features
  • Logistic regression using L2 regularization (scikit-learn default)