Skip to Main Content
This paper presents a novel method of improving the performance of a support vector machine (SVM) classifier by modifying kernel function. This is based on the differential approximation of metric. The method is to enlarge margin around the separating hyper-plane by modifying the kernel functions using a positive scalar function. Therefore, the separability is increased. Example is given specifically for modifying Gaussian Radial Basis Function kernel. Simulation results for both artificial and real data show remarkable improvement of generalization error and computational cost.