A Least Squares Support Vector Machine Sparseness Algorithm

Abstract This paper proposes a method which using density index function to sparse LS-SVM in highdimensional feature space, and gives a new method which takes each sample point as a clustering center to make hypersphere, so as to determine the fuzzy membership function in high-dimensional feature space, thus to establish a new fuzzy least squares support vector machine model, So it is different from the traditional pruning method which is in original space only. And the simulation experiment shows that the model is more effective and feasible.