🏦 Loan Approval Prediction using Neural Networks

A deep learning–based system to predict loan approval decisions using applicant financial and demographic data.
Developed and maintained by Surajit.

---

🚀 Overview

This project uses a PyTorch feed-forward neural network to predict whether a loan application will be Approved or Rejected, based on various financial and credit-related features.

The workflow:

1. Load a trained model and preprocessing objects (scaler and label encoders).
2. Read applicant data from an Excel file (loan_applications.xlsx).
3. Preprocess the data (encode categorical columns and scale numerical features).
4. Generate predictions and save the results in loan_applications_result.xlsx.

---

🧠 Model Architecture

The neural network (LoanNN) consists of:

• Input Layer → input_dim = 11
• Hidden Layer 1 → 64 neurons + ReLU + Dropout(0.3)
• Hidden Layer 2 → 32 neurons + ReLU
• Output Layer → 1 neuron + Sigmoid activation

This structure outputs a probability score representing the likelihood of loan approval.

---

📁 Project Structure

LOAN_APPROVAL/
├── .dist/                         # Distribution or build artifacts
├── data/                          # Raw or processed data files
├── Loan/                          # Core loan-related modules
├── models/                        # Trained models and architectures
├── output_data/                   # Generated results or outputs
│
├── .gitignore                     # Git ignore file
├── loan_approval.ipynb            # Main notebook for model training/evaluation
├── nn_model.ipynb                 # Neural network model development notebook
├── nn_run_model.py                # Script to run the trained neural network model
├── run_model.py                   # Prediction or model execution script
├── xcel_file_test.py              # Excel file testing and validation script
│
├── requirements.txt               # List of dependencies
└── README.md                      # Project documentation
                      # Project documentation

---

⚙️ Setup Instructions

1. Clone the repository

git clone https://github.com/Surajit09636/loan-approval-predictor.git
cd loan-approval-predictor

2. Create a virtual environment (recommended)

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

If you don't have requirements.txt yet, generate it using:

pip freeze > requirements.txt

---

📊 Input Data Format

The model expects an Excel file (loan_applications.xlsx) with the following columns:

Column	Description
dependents	Number of dependents
education	Graduate / Not Graduate
self_employed	Yes / No
imcome_annum	Annual income
loan_amount	Loan amount requested
loan_term	Term in months
cibil_score	Credit score
residential_assets_value	Value of residential assets
commercial_assets_value	Value of commercial assets
luxury_assets_value	Value of luxury assets
bank_asset_value	Value of bank assets

Example:

dependents	education	self_employed	imcome_annum	loan_amount	loan_term	cibil_score	residential_assets_value	commercial_assets_value	luxury_assets_value	bank_asset_value
1	Graduate	No	600000	150000	24	720	250000	50000	20000	70000

---

🧩 Running Predictions

Run the main script:

python xcel_file_test.py

It will:

• Load your model (./data/loan_nn_model_gpu.pth)
• Process loan_applications.xlsx
• Generate approval predictions
• Save results to loan_applications_result.xlsx

Sample output:

✅ Predictions saved to 'loan_applications_result.xlsx'

---

🧾 Output Format

The output Excel file will include:

• Prediction_Probability → Model confidence score
• Predicted_Status → “Approved” or “Rejected”

---

🧰 Dependencies

Main Python libraries:

• torch
• pandas
• scikit-learn
• openpyxl
• numpy

You can install them with:

pip install torch pandas scikit-learn openpyxl numpy

---

👨‍💻 Developed By

Surajit
📧 Email: surajitsutradhar010@gmail.com
💼 GitHub: https://github.com/Surajit09636

"Empowering smarter financial decisions through machine learning."

---

📝 License

This project is released under the MIT License.
You’re free to use, modify, and distribute with proper credit to Surajit.

---