Skip to content

openMF/scorecard-ai

Repository files navigation

Scorecard-AI 🧠📊

Hosted on HuggingFace

Scorecard-AI is a powerful tool designed for predicting Interest Repaid Derived and Nominal Interest Rates using machine learning models. This tool allows users to interactively select features, input their values, and choose the model to make predictions. If some feature values are not available, the tool replaces them with the average values or allows users to choose from predefined sets of values.

Features 🚀

  • User-friendly interface: Select features, input values, and choose models for prediction.
  • Model flexibility: Multiple model options to choose from, ensuring high customization.
  • Smart defaults: Missing feature values are replaced with average values automatically, saving time for users.
  • Predefined values: Users can also select from preset feature values.

Project Structure 🗂️

Web/
│
├── .gitattributes
├── README.md
├── dataset_link
├── interest_repaid_derived.py          # Original code for Interest Repaid Derived prediction
├── intrest_repaid_frontend.ipynb       # Modified Jupyter Notebook for the frontend
├── nominal_frontend.ipynb              # Modified Jupyter Notebook for Nominal Interest Rate frontend
└── nominal_interest_rate.py            # Original code for Nominal Interest Rate prediction

Web Folder 🌐

The Web folder contains the Django application used to host the Scorecard-AI tool on HuggingFace Docker Spaces. The frontend is built using HTML and CSS, providing a seamless user experience for interacting with the machine learning models.

Check out the live demo here: Scorecard-AI Live

Notebooks and Scripts 📒

  • interest_repaid_derived.py: Original Python script for predicting Interest Repaid Derived.
  • nominal_interest_rate.py: Original Python script for predicting Nominal Interest Rate.
  • intrest_repaid_frontend.ipynb & nominal_frontend.ipynb: These Jupyter notebooks are modified versions of the original Python scripts, adapted for the frontend application.

Usage 🛠️

  1. Clone the Repository:

    git clone https://github.com/yourusername/scorecard-ai.git
  2. Navigate to the Web Folder:: After cloning the repository, change the directory to the Web folder:

    cd scorecard-ai/Web
  3. Install Dependencies: Install the required Python packages by running:

    pip install -r requirements.txt
  4. Run Locally: You can run the Django application locally by navigating to the Web folder and running the following:

    python manage.py runserver
  5. Predict Interest Rates:

    • Select features via the frontend.
    • Choose the model you'd like to use for prediction.
    • Input values or allow the system to use average values where data is missing.

How it Works 🧠

  1. Feature Selection: Users are prompted to select the relevant features from the dataset for the prediction task.
  2. Input Values: Users input the values for the selected features. If any values are missing, the system automatically fills them with the average values from the dataset.
  3. Model Selection: The user selects the model they want to use for prediction. Each model is pre-trained and fine-tuned for either Interest Repaid Derived or Nominal Interest Rate prediction.
  4. Results: The predicted value is displayed in the frontend after processing.

Dependencies 📦

  • Django: Backend web framework.
  • Jupyter Notebook: Used for model development and testing.
  • Python: The primary language for the application.
  • HuggingFace Docker Spaces: The application is hosted using HuggingFace's Docker spaces.

Future Enhancements 🌱

  • Explore the other pre-processing techniques and models.
  • Explore other featureas we can look to predict.
  • Improve UI/UX for an even smoother experience.

Contributing 🤝

Contributions are welcome! Feel free to open an issue or submit a pull request. Be sure to follow the contribution guidelines.


References 🔗


Made with ❤️ by Parth Kaushal