Recently, sample complexity bounds have been derived for problems in(cid:173) volving linear functions such as neural networks and support vector ma(cid:173) chines. In this paper, we extend some theoretical results in this area by deriving dimensional independent covering number bounds for regular(cid:173) ized linear functions under certain regularization conditions. We show that such bounds lead to a class of new methods for training linear clas(cid:173) sifiers with similar theoretical advantages of the support vector machine. Furthermore, we also present a theoretical analysis for these new meth(cid:173) ods from the asymptotic statistical point of view. This technique provides better description for large sample behaviors of these algorithms.

Recently, sample complexity bounds have been derived for problems involving linear functions such as neural networks and support vector machines. In this paper, we extend some theoretical results in this area by deriving dimensional independent covering number bounds for regularized linear functions under certain regularization conditions. We show that such bounds lead to a class of new methods for training linear classifiers with similar theoretical advantages of the support vector machine. Furthermore, we also present a theoretical analysis for these new methods from the asymptotic statistical point of view. This technique provides better description for large sample behaviors of these algorithms.

Recently, sample complexity bounds have been derived for problems involving linear functions such as neural networks and support vector machines. In this paper, we extend some theoretical results in this area by deriving dimensional independent covering number bounds for regularized linear functions under certain regularization conditions. We show that such bounds lead to a class of new methods for training linear classifiers with similar theoretical advantages of the support vector machine. Furthermore, we also present a theoretical analysis for these new methods from the asymptotic statistical point of view. This technique provides better description for large sample behaviors of these algorithms.

Recently, sample complexity bounds have been derived for problems involving linearfunctions such as neural networks and support vector machines. In this paper, we extend some theoretical results in this area by deriving dimensional independent covering number bounds for regularized linearfunctions under certain regularization conditions. We show that such bounds lead to a class of new methods for training linear classifiers withsimilar theoretical advantages of the support vector machine. Furthermore, we also present a theoretical analysis for these new methods fromthe asymptotic statistical point of view. This technique provides better description for large sample behaviors of these algorithms. 1 Introduction In this paper, we are interested in the generalization performance of linear classifiers obtained fromcertain algorithms.