Abstract:
Abstract. The Support Vector Machine (SVM) has emerged in recent years as a popular approach to the classi¿cation of data. One problem that faces the user of an SVM is how to choose a kernel and the speci¿c parameters for that kernel.Applications of an SVM therefore require a search for the optimum settings for aparticular problem. This paper proposes a classi¿cation technique, which we call the Genetic Kernel SVM (GK SVM), that uses Genetic Programming to evolve akernel for a SVMclassi¿er. Results of initial experiments with the proposed tech-nique are presented. These results are compared with those of a standard SVM classi¿er using the Polynomial or RBF kernel with various parameter settings.
