🧠 Neural Cosine Approximator

Approximate the cosine function using a simple neural network from scratch!

This project demonstrates how a neural network can learn to approximate the mathematical cos(x) function using supervised learning.

🚀 Features

• 📚 Implemented entirely in Python and Jupyter Notebook.
• 🧮 Uses a simple fully connected neural network.
• 🏗️ Built from scratch — no deep learning frameworks (such as TensorFlow or PyTorch) required!
• 📈 Visualizes both the training process and the approximation performance.
• 🎓 Great for educational purposes and understanding how neural networks learn mathematical functions.

📝 Project Structure

• Neural_Cosine_Approximator.ipynb — Main notebook containing:
  • Data preparation
  • Neural network implementation
  • Training loop
  • Evaluation & visualization

📦 Dependencies

• Python >= 3.8
• NumPy
• Matplotlib
• Jupyter Notebook

You can install the required libraries using:

pip install numpy matplotlib jupyter

🏃‍♂️ Usage

1. Clone the repository:

git clone https://github.com/YOUR_USERNAME/YOUR_REPOSITORY.git
cd YOUR_REPOSITORY

2. Launch the notebook:

jupyter notebook Neural_Cosine_Approximator.ipynb

3. Run the cells and enjoy watching your neural network approximate cos(x)!

📚 Learning Goals

• Understand how neural networks can approximate continuous functions.
• Learn how to implement a basic neural network without external frameworks.
• Visualize model learning and performance.

📜 License

This project is licensed under the MIT License.

✨ Author

Sayyed Hossein Hosseini DolatAbadi

---

🌟 If you like this project, consider giving it a ⭐ on GitHub!