Chapter 11: Support vector machine
Daniela Calvetti, Erkki Somersalo
Abstract
Binary classification of data is one of the most ubiquitous operations, arising in a wide variety of applications. The design of a binary classifier typically consists of finding a rule that partitions data points in two clusters so as to minimize classification error when tested on a set of annotated data. The support vector machine (SVM) algorithm, in particular its generalized kernel-based extension, has gained a lot of popularity as a binary classifiers. In this chapter we outline the mathematical principles leading to the SVM algorithm and discuss some aspects related to its implementation, without going too deeply into questions of optimization methods that are not a central topic in this book.
Topics & Concepts
Support vector machineComputer scienceStructured support vector machineBinary classificationArtificial intelligenceBinary numberMachine learningExtension (predicate logic)PopularityData miningRelevance vector machineClassifier (UML)Kernel methodKernel (algebra)Pattern recognition (psychology)MathematicsArithmeticProgramming languagePsychologyCombinatoricsSocial psychologyAdvanced Clustering Algorithms ResearchAdvanced Algorithms and ApplicationsFace and Expression Recognition