---

Svm

Compounds
class	SvmData
class	SvmLinsys
class	SvmResiduals
class	SvmVars

---

Detailed Description

Solve the QP arising in support vector machine with a linear kernel.

Input is a list of points in N-space, each point labelled with one of two possible labels. Output is a hyperplane that separates the two labelled sets of points, if possible, or "nearly separates" them according to minimizing an objective function in which violations (distance of points on the wrong side from the plane) are weighed against a norm of the hyperplane gradient.

For further information on this formulation, see Chapter 5 of V. Vapnik: "The Nature of Statistical Learning Theory", 2nd edition, Springer, 1999.

---