Credit Risk Prediction System

📌 Overview

The Credit Risk Prediction System is a machine learning application that evaluates a person's creditworthiness based on financial and personal details. It utilizes Apache Spark for scalable data processing and Streamlit for an interactive user interface.

This project consists of two main components:

1. Model Training (credit_risk_model_trainer.py)

   • Trains a Random Forest classifier on the German Credit Risk Dataset.
   • Performs categorical encoding, feature engineering, and data balancing to improve performance.
   • Saves the trained model and feature indexers for future predictions.

2. Prediction Application (credit_risk_detector.py)

   • A Streamlit-powered UI where users input financial details.
   • Loads the trained model and predicts Good or Bad credit risk.
   • Displays confidence levels for each prediction.

---

🚀 Features

✔ End-to-End Machine Learning Pipeline
✔ Scalable Processing with Apache Spark
✔ Interactive Credit Risk Prediction via Streamlit UI
✔ Feature Engineering & Class Imbalance Handling
✔ Model Performance Evaluation (ROC-AUC Score)

---

🏗️ Installation & Setup

Prerequisites

Ensure you have the following installed:

• Python (>= 3.8)
• Apache Spark (>= 3.0)
• Streamlit (pip install streamlit)
• PySpark (pip install pyspark)

Clone the Repository

git clone https://github.com/yourusername/credit-risk-prediction.git
cd credit-risk-prediction

Install Dependencies

pip install -r requirements.txt

Download the Dataset

The model training script automatically downloads the dataset if it's missing. However, you can manually download it from:
German Credit Dataset - UCI

---

⚙️ How to Run

1️⃣ Train the Model

Run the following command to train the credit risk model:

python credit_risk_model_trainer.py

This will:

• Download the dataset (if not present)
• Train a Random Forest Classifier
• Save the trained model in ./credit_risk_model/
• Store feature indexers in ./indexers/

2️⃣ Start the Credit Risk Prediction App

Once the model is trained, launch the Streamlit app:

streamlit run credit_risk_detector.py

---

🎯 Usage Guide

• Open the Streamlit UI in your browser.
• Enter the applicant’s financial and demographic details.
• Click Predict Credit Risk to classify the applicant as Good ✅ or Bad ❌.
• View the prediction confidence score.
• Check Feature Descriptions for input details.

---

📊 Model Performance

After training, the model evaluates performance using ROC-AUC Score:

Random Forest ROC-AUC Score: 0.85 (Example)

Additionally, feature importance is displayed to highlight key predictive variables.

---

🏗️ Project Structure

📂 credit-risk-prediction
 ├── 📜 credit_risk_model_trainer.py  # Model training script
 ├── 📜 credit_risk_detector.py       # Streamlit UI for credit risk prediction
 ├── 📂 credit_risk_model/            # Trained Random Forest Model
 ├── 📂 indexers/                      # Categorical feature encoders
 ├── 📄 requirements.txt               # Python dependencies
 ├── 📄 README.md                      # Project documentation

---

🤝 Contribution

We welcome contributions!
Feel free to fork this repo, create a pull request, or report issues.

---

📄 License

This project is licensed under the MIT License.

---