Loan Eligibility Predictor

An application for prediction loan eligibility for applicants based on their personal, financial, and property-related information.

---

Table of Contents

• Overview
• Features
• Installation
  • Using pip
  • Using Conda
• Usage
• License

---

Overview

The Loan Eligibility Predictor is a tool built to assist banks and financial institutions in determining whether an applicant qualifies for a loan. Trained on a Random Forest Classifier, which is then deployed as a web application using Streamlit. The app allows users to input their details and receive instant feedback on their loan eligibility, along with a confidence score indicating the model's certainty.

---

Features

• Predicts loan eligibility in real-time.
• Provides a confidence score for each prediction.
• User-friendly interface powered by Streamlit.
• Support for both neural network and random forest classifier models.

---

Project Structure

LoanEligibilityPredictor/
├── assets/                 # Contains images and other assets
├── dataset/                # Contains the dataset files
├── app.py                  # Streamlit web application
├── eda.ipynb               # Exploratory Data Analysis notebook
├── LoanerNN.ipynb          # Neural Network model notebook
├── LoanerRFC.ipynb         # Random Forest Classifier notebook
├── nn.py                   # Neural Network model script
├── utils.py                # Utility functions
├── requirements.txt        # Python dependencies
├── environment.yml         # Conda environment configuration
├── README.md               # Project documentation
└── LICENSE                 # License information

Installation

You can install and run this application using either pip or conda.

Using pip

1. Clone the Repository:

   git clone https://github.com/SamarthPyati/LoanEligibilityPredictor.git
   cd loan-eligibility-predictor

2. Set Up a Virtual Environment (Optional but Recommended):

   python3 -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install Dependencies:

   pip install -r requirements.txt

Using Conda

1. Clone the Repository:

   git clone https://github.com/SamarthPyati/LoanEligibilityPredictor.git
   cd loan-eligibility-predictor

2. Create and Activate a New Conda Environment:

   conda env create -f environment.yml
   conda activate loan-predictor

3. Alternative: Manual Conda Setup (if environment.yml is not working):

   conda create -n loan-predictor 
   conda activate loan-predictor
   pip install -r requirements.txt

---

Usage

1. Run the Streamlit App:

   streamlit run app.py

2. Enter Applicant Information: Fill in the required fields with the applicant's details.

3. Get Prediction: Click the "Predict" button to receive the loan eligibility prediction.

---

Training the Model

---

# Run the training notebook
jupyter notebook LoanerRFC.ipynb

Alternatively, for the Neural Network model:

# Run the training notebook
jupyter notebook LoanerNN.ipynb

These notebooks will guide you through the training process, including data preprocessing, model training, and evaluation.

Testing the Model

Testing is integrated within the training notebooks. After training, the notebooks include evaluation metrics such as accuracy, precision, recall, and ROC-AUC scores to assess model performance.

Running the Web Application

To launch the Streamlit web application for real-time loan eligibility predictions:

streamlit run app.py

This will open a web interface where you can input applicant details and receive eligibility predictions.

File Descriptions

• app.py: Main Streamlit application script that provides a web interface for users to input data and receive predictions.
• eda.ipynb: Notebook for Exploratory Data Analysis, including data visualization and initial insights.
• LoanerRFC.ipynb: Notebook detailing the training process using the Random Forest Classifier.
• LoanerNN.ipynb: Notebook detailing the training process using a Neural Network model.
• nn.py: Python script containing the implementation of the Neural Network model.
• utils.py: Contains utility functions used across the project, such as data preprocessing functions.
• requirements.txt: Lists all Python dependencies required to run the project.
• environment.yml: Conda environment configuration file listing dependencies and environment settings.
• assets/: Directory containing images and other assets used in the project.
• dataset/: Directory containing the dataset files used for training and testing.

License

This project is licensed under the MIT License. See the LICENSE file for details.