Textbook of AI

Abstract. Introduces the soft-margin support vector machine, extending the maximum-margin classifier to handle non-separable data via slack variables. Combined with the kernel trick, SVMs became the dominant classifier of the late 1990s and 2000s.