Beginning Anomaly Detection Using

support vector is a vector parallel to the hyperplane that acts as the 

decision boundary, containing a point that is closest to the 

hyperplane, and helps 

establish a margin for the decision boundary. In this example, the hyperplane is a line 

because there are only two dimensions. In three dimensions, the hyperplane would be a 

plane, and in four dimensions, it would be a three-dimensional space, and so on.

The most optimal hyperplane would involve the support vectors establishing a 

maximum margin for the hyperplane. The example in Figure 

2-32

 is not optimal, so let’s 

look for a more optimal hyperplane in Figure 

2-33

.