A Neural Network that predicts handwritten digits using the MNIST dataset. The project involves data preprocessing, model development, and evaluation using Python and Jupyter Notebook.
This project focuses on building and training a Neural Network to predict handwritten digits. By utilizing the popular MNIST dataset and implementing advanced techniques like Mini-batch gradient descent, the project demonstrates a high level of accuracy in digit recognition.
- Utilized the Python programming language and Jupyter Notebook for seamless data preprocessing, model development, and evaluation.
- Implemented the Mini-batch gradient descent technique to optimize model training and enhance convergence speed.
- Achieved an impressive accuracy rate of 96.63% on the test dataset, showcasing the robustness of the developed model.
- Gained practical experience and developed a deeper understanding of machine learning and data science through hands-on involvement with the project.
- Python
- Jupyter Notebook
- Clone this repository:
git clone https://github.com/alex8430/Neural-Network-From-Scratch.git - Navigate to the project directory:
cd Neural-Network-From-Scratch - Install required packages:
pip install -r requirements.txt
- Open
handwritten_digit_recognition.ipynbin Jupyter Notebook. - Follow the step-by-step instructions to preprocess data, develop the model, and evaluate its performance.
- Experiment with different parameters and techniques to further enhance the model's accuracy.
We welcome contributions to improve the project. To contribute:
- Fork this repository.
- Create a new branch:
git checkout -b feature-xyz - Make your changes and commit them:
git commit -m "Add feature" - Push to the branch:
git push origin feature-xyz - Open a pull request.
For questions or suggestions, feel free to reach out to me at pankajvermacr7@gmail.com or through LinkedIn.