Please use a desktop to appropriately visualize this page.

**Perceptrons** are supervised learning models used to classify data into binary classes. They are one of the simplest models around, and thus serve as a good introduction to machine learning.

This page contains a running visualization of the Perceptron Learning Algorithm (PLA). First, a target function is generated randomly, and then, a set of observations is uniformly generated to populate the dataset. The learning algorithm is executed according to the lines below.

Number of points:

Generate linearly separable data

Generate linearly separable data

0. Initialize \( \mathbf{w} \leftarrow \mathbf{0} \)

1. While there are misclassified points:

1.1. Pick a misclassified point \(\mathbf{x}_n\)

1.2. Update weights: \( \mathbf{w} \leftarrow \mathbf{w} + y_n \mathbf{x}_n \)

Iterations: 0

Misclassifications: 0

Misclassifications: 0