Machine Learning for the Web

Support vector machine

This algorithm, Support Vector Machine (SVM), tries to geometrically separate the dataset into two subsets labeled with y_i=+1 and y_i=-1. The next figure shows the data perfectly separated into two classes (empty circles and black circles), that is, the case the data in which the decision boundary (or hyperplane) given by the black line fully separates the two classes (in other words, there are no misclassified data points):

Sketch of the dataset separated into two classes (empty and filled circles) by the black line (decision ...