What is the term for a decision boundary in support vector machines that has parallel and equidistant lines on either side?

The term "hyperplanes" refers to the concept used in support vector machines (SVM) to describe the decision boundary that separates different classes in the feature space. In SVM, hyperplanes are defined as high-dimensional planes that can separate data points belonging to different categories.

In a two-dimensional space, this would manifest as a line, whereas in a higher-dimensional space, it extends to more complex structures. The critical aspect of the hyperplane in the context of SVM is that it is flanked by parallel and equidistant margins that represent the support vectors—the closest data points from each class. These margins are essential for maximizing the margin of separation between classes, which is a fundamental goal of SVM.

Understanding this concept is vital for grasping how SVM models decide which hyperplanes to use for classification and how those hyperplanes can effectively discriminate between classes based on the training data.