\magnification=1200
\baselineskip=20pt
\nopagenumbers
\font\big=cmr12 scaled \magstep2

\centerline{\bf STANFORD UNIVERSITY}
\centerline{\bf DEPARTMENT OF STATISTICS}
\centerline{\big DEPARTMENTAL SEMINAR}
\bigskip

\baselineskip=12pt
\centerline{4:15 p.m., Tuesday, August 14, 2001}
\centerline{Sequoia Hall Rm. 200}
\centerline{(Cookies at 3:45 in 1st Floor Lounge)}

\bigskip
\baselineskip=15pt
\centerline{\sl Yi Lin}
\centerline{\sl Department of Statistics}
\centerline{\sl University of Wisconsin}
\centerline{\sl Madison}
\bigskip
\centerline{\bf The support vector machine and related classification methods}

\bigskip

In this talk, I will present some results on the statistical
properties of the support vector machine and some related
classification methods. I will first give a (very brief) overview of
the development of the support vector machine. After introducing a
regularization formulation of the support vector machine, I will
present some results shedding light on how the support vector machine
works. Consistency and rate of convergence to the Bayes risk (the
expected misclassification rate of the theoretical optimal
classification rule) will be introduced.

Many machine learning methods for classification are based on cost
functions that depend only on the margin. The support vector machine
is one example. In this talk I will also present some results on
methods based on a general class of cost functions of the margin.

\bye