This repository contains the implementation of a simple Perceptron neural network. The Perceptron is a foundational concept in machine learning and serves as the building block for more complex neural networks.
The purpose of this project is to demonstrate the basic workings of a Perceptron neural network using synthetic data. The notebook walks through the steps of data generation, model definition, training, and visualization of the decision boundary.
The notebook is divided into several sections:
-
Data Generation: Creates a synthetic dataset with two features and two classes.
-
Data Preview: Displays a sample of the generated data and their corresponding labels.
-
Label Conversion: Converts the labels from {0, 1} to {-1, 1} to be compatible with the Perceptron.
-
Sign Function: Defines a helper function to return the sign of a value.
-
Perceptron Model: Implements the Perceptron model class, which includes weight initialization and the forward pass.
-
Optimizer: Implements an optimizer to update the model weights and bias.
-
Training and Visualization: Trains the Perceptron model and visualizes the decision boundary along with the data points.
The notebook outputs a plot showing the data points and the decision boundary learned by the Perceptron. This visualization helps to understand how the Perceptron classifies the data.
Here is a sample visualization of decision boundary:
To run this project, you'll need to have Python installed along with the following libraries:
numpymatplotlibscikit-learn
You can install these dependencies using pip:
pip install numpy matplotlib scikit-learnClone the repository and navigate to the project directory:
git clone https://github.com/yourusername/perceptron-neural-network.git
cd perceptron-neural-networkOpen the Jupyter notebook:
jupyter notebook Perceptron_Neural_Network.ipynbRun all cells to generate the synthetic data, train the Perceptron model, and visualize the results.