支持向量机及文本分类.pptVIP

Support Vector Machines outline Linear SVM - Separable Case Hyperplane classifiers g(x) = w?x+b Usually we add the constraint |g(x)|= | w?x+b |= 1 That is, all the training examples satisfy: equivalently: Let d+ (d-) be the shortest distance from the hyperplane to the closest positive (negative) example. The margin of the hyperplane is defined to be d+ + d- separating hyperplane: w?x+b = 0 decision function: f(x) = sgn(w?x+b) The margin is given by dual problem maximize : subject to: ?i?0 and Linear SVM - Non-Separable Case Kernel function Text Categorization PROPERTIE