A Hybrid Wavelet Kernel Construction for Support Vector Machine Classification

Support Vector Machine (SVM) is a machinelearning algorithm, which learns to perform the classification task through a supervised learning procedure, based on preclassified data examples. SVM uses a device called kernel mapping to map the non-linear data in input space to a higher dimensional feature space where the data becomes linearly separable. A hybrid wavelet kernel construction for support vector machine is introduced in this paper. The construction involves a multi-dimensional orthogonal sinc wavelet function together with one of the conventional kernel functions. We show that the hybrid kernel is an admissible kernel. Hybrid kernels provide better classification of the signal points in the mapped feature space. The hybrid kernel thus constructed is used for the classification of Cardiac Single Photon Emission Computed Tomography (SPECT) images and Cardiac Arrhythmia signals. The experimental results show that promising generalization performance can be achieved with the hybrid kernel, compared to conventional kernels.