This project focuses on letter recognition using a custom-built feedforward neural network. It classifies binary images of letters A, B, and C using NumPy for computation and Matplotlib for visualizations. From scratch, the network implements forward propagation, sigmoid activation, and custom backpropagation, offering insights into the workings of neural networks.
- 🏗️ Built a feedforward neural network entirely from scratch using NumPy.
- 🔁 Implemented custom backpropagation and sigmoid activation without any ML libraries.
- 🧮 Trained on binary image datasets of letters A, B, and C.
- 📊 Tracked training accuracy over epochs to evaluate performance.
- 📉 Visualized model predictions and training results using Matplotlib.
- Python
- NumPy
- Matplotlib
- Object-Oriented Programming
- Clone the repository:
git clone https://github.com/SarthakKumarPathak/neural-network-classifier.git
- Install dependencies: (All required libraries are standard in most Python environments): pip install numpy matplotlib
- Run the training script: python train_classifier.py
Achieved high classification accuracy on the letter dataset.
Visualized training performance with plots of accuracy vs. epochs.
Displayed predictions vs. actual labels to showcase classifier performance.
This project is licensed under the MIT License - see the LICENSE file for details.
Made with ❤️ by Sarthak Kumar Pathak