Abstract
The sparse model character of 1-norm penalty term of Least Absolute Shrinkage and Selection Operator (LASSO) can be applied to automatic feature selection. Since 1-norm SVM is also designed with 1-norm (LASSO) penalty term, this study labels it as LASSO for classification. This paper introduces the smooth technique into 1-norm SVM and calls it smooth LASSO for classification (SLASSO) to provide simultaneous classification and feature selection. In the experiments, we compare SLASSO with other approaches of "wrapper" and "filter" models for feature selection. Results showed that SLASSO has slightly better accuracy than other approaches with the desirable ability of feature suppression. © 2010 IEEE.