Two enhancements are proposed to the application and theory of support vect
or machines. The first is a method of multicategory classification based on
the binary classification version of the support vector machine (SVM). The
method, which is called the M-ary SVM, represents each category in binary
format, and to each bit of that representation is assigned a conventional S
VM. This approach requires only [log(2)(K)] SVMs, where K is the number of
classes. We give an example of classification on an octaphase-shift-keying
(8-PSK) pattern space to illustrate main concepts.
The second enhancement is that of adding equality constraints to the conven
tional binary classification SVM. This allows pinning the classification bo
undary to points that are known a priori to lie on the boundary. Applicatio
ns of this method often arise in problems having some type of symmetry. We
present one such example where the M-ary SVM is used to classify symbols of
a two-user, multiuser detection pattern space.