https://github.com/nishumehta/loan-default-prediction

A machine learning web app to predict the likelihood of a loan default based on applicant data. Trained Logistic Regression, Random Forest, and XGBoost models. Built with Python, Streamlit, and deployed on Streamlit Cloud.
https://github.com/nishumehta/loan-default-prediction

logistic-regression machine-learning python random-forest streamlit web-app xgboost

Last synced: about 2 months ago
JSON representation

A machine learning web app to predict the likelihood of a loan default based on applicant data. Trained Logistic Regression, Random Forest, and XGBoost models. Built with Python, Streamlit, and deployed on Streamlit Cloud.

• Host: GitHub
• URL: https://github.com/nishumehta/loan-default-prediction
• Owner: NishuMehta
• Created: 2025-07-04T16:56:21.000Z (12 months ago)
• Default Branch: main
• Last Pushed: 2025-07-08T04:22:11.000Z (12 months ago)
• Last Synced: 2025-09-24T16:59:28.124Z (9 months ago)
• Topics: logistic-regression, machine-learning, python, random-forest, streamlit, web-app, xgboost
• Language: Jupyter Notebook
• Homepage: https://loan-default-prediction-nishu.streamlit.app/
• Size: 8.2 MB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0

Awesome Lists containing this project

README

          
# Loan Default Risk Predictor

This project predicts whether a loan applicant is likely to **default** using machine learning. It provides a user-friendly **Streamlit web app** where users can input applicant data and receive a risk prediction instantly.

##  Live App

[Click here to try the app](https://loan-default-prediction-nishu.streamlit.app/)  

![](/images/Demo_1.png)

![](/images/Demo_2.png)

---

## Features

- **24 input fields** based on user profile and financial history

- Clean interface with two-column input layout

- Prediction label: `Default` or `No Default`

- Confidence score from logistic regression

---

## Dataset Overview

Each row in the dataset contains features like:

- `Age`, `Income`, `LoanAmount`, `CreditScore`, `DTIRatio`

- Binary indicators: `HasMortgage`, `HasDependents`, `HasCoSigner`

- Encoded categorical features: `Education`, `EmploymentType`, `MaritalStatus`, `LoanPurpose`

---

## Model Training

Three models were trained:

| Model               | ROC-AUC Score |

|--------------------|---------------|

| Logistic Regression| **0.7531** (Best) |

| Random Forest       | 0.7390        |

| XGBoost             | 0.7409        |

The final deployed model is **Logistic Regression** due to its balance of accuracy, interpretability, and low overfitting.

---

## How to Run Locally

```bash

git clone https://github.com/NishuMehta/Loan-Default-Prediction.git

cd Loan-Default-Prediction

pip install -r requirements.txt

streamlit run app/main.py

## Author

- [Nishu Mehta](https://github.com/NishuMehta)

## Project Link

 [GitHub Repository](https://github.com/NishuMehta/Loan-Default-Prediction)