A New Features Selection model: Least Squares Support Vector Machine with Mixture of Kernel

In this paper, a least squares support vector machine with mixture kernel (LS-SVM-MK) is proposed to solve the problem of the traditional LS-SVM model, such as the loss of sparseness and robustness. Thus that will result in slow testing speed and poor generalization performance. The revision model LS-SVM-MK is equivalent to solve a linear equation set with deficient rank just like the over complete problem in independent component analysis. A minimum of 1-penalty based object function is chosen to get the sparse and robust solution. Some UCI datasets are used to demonstrate the effectiveness of this model. The experimental results show that LS-SVM-MK can obtain a small number of features and improve the generalization ability of LS-SVM.