2.3.2. Support Vector Machine (SVM) Algorithm
A supervised learning technique that generates input-output mapping
functions from a collection of labeled training data is the
Support Vector Machine (SVM). The
mapping function may be a function of classification or a function of
regression. Nonlinear kernel functions are also used for classification
to translate input data to a high-dimensional feature space where the
input data becomes more separable than the original input space.
Hyperplanes of maximum-margin are then formed. Only a subset of the
training data near the class boundaries depends on the generated model.
In the present study, kernel type was radial basis polynomial,
regularization parameter (C) was ten, gamma was 0.1, and stopping
criteria was 1.0E (-3). The
hyperparameters of the SVM algorithm was tune using a modified
sequential minimal optimization (SMO) method (11, 13).