Learn

Perceptron

Finding a Linear Classifier

Let’s recap what you just learned!

The perceptron has inputs, weights, and an output. The weights are parameters that define the perceptron and they can be used to represent a line. In other words, the perceptron can be visualized as a line.

What does it mean for the perceptron to correctly classify every point in the training set?

Theoretically, it means that the perceptron predicted every label correctly.

Visually, it means that the perceptron found a *linear classifier*, or a *decision boundary*, that separates the two distinct set of points in the training set.

In the plot on the right, you should be able to see the *linear classifier* that was found by the perceptron in the last iteration of the training process.