Скачать или смотреть Perceptron Hyperplane Visualization

This video uses the perceptron learning algorithm to classify a set of data.

The data points hovering in the middle are the the training data set--their colors (light blue and red) are representing the true target binary value. The points along the surface of the hyperplane are mapped with their target value included, thus +1 for light blue and -1 for red. The decision boundary eventually converges since the data is linearly separable.

After the perceptron is trained, the perceptron is tested with a new set of data outside the space of the points training points. These new points appear after the decision boundary has converged, and are marked in bold. Any misclassified points appear in pink, which in this case there were none.