Loan Default Risk Prediction using Machine Learning

Overview

This project is focused on building a classification model that predicts whether a loan applicant will be approved or not based on key demographic and financial attributes. The goal is to assist financial institutions in automating the screening process and improving decision accuracy.

Dataset

• Source: Kaggle - Loan Prediction Dataset
• Files Used:
  • train.csv — 614 rows of labeled training data
  • test.csv — unlabeled data for prediction
  • loan_predictions.csv — model-generated predictions

Tools & Technologies

• Python (Pandas, NumPy, Matplotlib, Seaborn)
• Scikit-learn (Logistic Regression, Random Forest)
• Jupyter Notebook
• Git, GitHub

Steps Performed

1. Data Cleaning & Preprocessing
2. Exploratory Data Analysis (EDA)
3. Feature Encoding
4. Model Training (Logistic Regression, Random Forest)
5. Model Evaluation (Accuracy, Precision, Recall, F1 Score)
6. Prediction on Test Data
7. Business Insights Summary

Results

• Best Model: Logistic Regression
• Accuracy: ~78.9%
• Key Influencing Features: Credit History, Applicant Income, Education
• Business Value: Streamlined loan approval process with minimal manual effort

How to Run

1. Clone this repo or download the ZIP
2. Open the notebook loan_default_prediction.ipynb in Jupyter or Colab
3. Run all cells sequentially to reproduce the results

Author

Prakash — Aspiring Data Analyst