Understanding and formulation of various kernel techniques for suport vector machines

Abstract

Support Vector Machines (SVM's) are supervised learning algorithms which can be used for analyzing patterns and classifying data. This supervised algorithm is applicable for binary class as well as multiclass classification. The core idea is to build a hyperplane which can easily separate the training examples. For binary class, SVM constructs a hyper-plane which can easily separate d-dimensional training examples perfectly into 2-classes. but sometimes, the training examples are not linearly separable. Thus, for non-linear training examples, SVM introduced Kernel functions which transforms the data into high dimensional space where the data can be separated linearly. For minimizing the test error and for improving classification accuracy, kernels functions are used. This paper explains applications of kernels in support vector machine and provide information about the properties of these kernels and situations in which they can be used.