Department of Statistics,
Title: A Bahadur Representation of the Linear Support Vector Machine
ABSTRACT: The support vector machine has been successful in a variety of applications. Also on the theoretical front, statistical properties of the support vector machine
have been studied quite extensively with a particular attention to its Bayes risk consistency under proper conditions. Bridging conventional statistics to machine learning,
some recent results on the asymptotic behavior of the coefficients of the linear support vector machine will be presented in this talk. In particular, a Bahadur-type
representation of the coefficients is established under appropriate conditions, and their asymptotic normality and statistical variability are derived. Implications of the
results on related statistical inferences will be discussed.