Neural Network from Scratch

This project implements a basic neural network from scratch using Python. The network can be trained to classify or predict based on input data without the use of external machine learning libraries.

Features

• Forward propagation
• Backpropagation
• Adjustable learning rate
• Multi-layer architecture

How to Run

1. Clone the repository:

git clone https://github.com/darwin-luque/neural-network.git
cd neural-network

2. Install dependencies:

pip install -r requirements.txt

3. Run the script:

python testing.py

Note: Generate your own dataset. Maybe try using the XOR problem to test the network.

How It Works

• Forward Propagation: The network computes the output by passing inputs through multiple layers.
• Backpropagation: The error is computed and propagated back to update weights.

Example

To showcase how the neural network works, you can run it on a simple dataset (e.g., XOR problem) and evaluate the model’s accuracy.

Future Improvements

• Adding more complex activation functions like ReLU or Tanh.
• Implementing optimization techniques like momentum or learning rate decay.
• Expanding the network to include convolutional neural networks (CNNs) or recurrent neural networks (RNNs).