Suppose we can draw a line on the plane, such that all points in class A lie
to one side and all points in line B lie to the other. Then, the line can be used to
classify new points, whose class we don’t already know. But there may be many
possible such lines that can separate points in class A from points in class B. A
few such lines are shown in Figure 20.5. The Support Vector Machine classifier
chooses the line whose distance from the nearest point in either class (from the
points in the training data set) is maximum. This line (called the maximum margin
line) is then used to classify other points into class A or B, depending on which
side of the line they lie on. In Figure 20.5, the maximum margin line is shown in
bold, while the other lines are shown as dashed lines