18
Classifiers: Support Vector Machine
§Relatively new
§Universal learner
–Basic form: linear
–Many different kernels: polynomial, radial basis function, neural nets, etc.
–
§Implementation:                  by T. Joachims
–Linear
–Radial Basis Function (RBF)
§
§
(1)
One important characteristic of Support Vector machines is that they are in some sense universal learners …in its basic form, SVM implements a linear classifier, however many different kernels have been adopted which can be used to implement other types of classifiers, including …The implementation we used is the …. And we tested both the linear and the radial basis function kernels.