A structural similarity kernel is presented in this paper for SVM learning, especially for learning with imbalanced datasets. Kernels in SVM are usually pairwise, comparing the sim...
In this paper, a novel method for reducing the runtime complexity of a support vector machine classifier is presented. The new training algorithm is fast and simple. This is achiev...
Abstract. The present study investigates a geometrical method for optimizing the kernel function of a support vector machine. The method is an improvement of the one proposed in [4...
— Among all ECG components, QRS complex is the most significant feature. Entropy based method for the detection of QRS complexes (cardiac beat) in the single lead Electrocardiogr...
An algorithmfor data condensation using support vector machines (SVM's)is presented. The algorithm extracts datapoints lying close to the class boundaries,whichform a much re...