\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, July 17, 2001}
\centerline{Sequoia Hall Rm. 200}
\centerline{(Cookies at 3:45 in 1st Floor Lounge)}

\bigskip
\baselineskip=15pt
\centerline{\sl Kesar Singh}
\centerline{\sl Department of Statistics}
\centerline{\sl Rutgers University}
\bigskip
\centerline{\bf Bootstrap based outlier detection plot (bootlier plot)}

\bigskip

The bootstrap density plot (histogram) of ``mean minus trimmed mean'' 
for a suitable trimming number is proposed as a plot to detect outliers 
in a data set. This plot is multimodal (bumpy) in the presence of
outliers. As a tool for outlier detection, this plot has some advantages
over the traditional method, which are pointed out in the paper. 
Extensions to multivariate outliers and outliers relative to a given
(nonlinear) statistic are considered. Some theoretical results are
presented which explain how the multimodality in the bootlier plot is
caused by outlier(s) in the sample.

\bye