Kernel Function Selection for the Solution of Classification Problems via Support Vector Machines

One of the most important machine learning algorithms developed for to accomplish classification task of data mining is Support Vector Machines. In the literature, Support Vector Machines has been shown to outperform many other techniques. Kernel function selection and parameter optimization play important role in implementation of Support Vector Machines. In this study, Kernel function selection process was ground on the randomized block experimental design. Univariate ANOVA was utilized for kernel function selection. As a result, the research proved that radial based Kernel function was the most successful Kernel function was proved.