Title:
The Marked Point Bootstrap

Abstract:
The bootstrap is a popular procedure for obtaining standard errors of
estimates. The original bootstrap was meant to be applied to
independent data, but various extensions to the bootstrap have been
developed for use with dependent data such as time series and spatial
data. Examples are subsampling and the block bootstrap. I will
introduce the marked point bootstrap as yet another method for
bootstrapping spatial data. It has a relatively simple formulation,
can be much faster than block bootstrap, and
has some desirable properties.