Data classification using Sparse and Robust model: least squares support vector machine with L1 norm
Liwei Wei1, Hao Yu2, Junhua Liu1
COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(12B) 686-691
1China National Institute of Standardization, No.4 Zhichun Road, Beijing, China
2China Agricultural University Library, No.2 Yuanmingyuan West Road, Beijing, China
Least squares support vector machine (LS-SVM) has an outstanding advantage of lower computational complexity than that of standard support vector machines. Its shortcomings are the loss of sparseness and robustness. Thus it usually results in slow testing speed and poor generalization performance. In this paper, a least squares support vector machine with L1 norm (LS-SVM-L1) is proposed to deal with above shortcomings. This method is equivalent to solve a linear equation set with deficient rank just like the over complete problem in independent component analysis (ICA). A minimum of 1-norm based object function is chosen to get the sparse and robust solution based on the idea of basis pursuit (BP) in the whole feasibility region. Some UCI datasets are used to demonstrate the effectiveness of this model. The experimental results show that LS-SVM-L1 can obtain a small number of support vector and improve the generalization ability of LS-SVM.