A new method for estimating the distribution of scan statistics for a two-dimensional Poisson process (Q1398016)

For a two-dimensional Poisson point process, its scan statistic is the largest number of points in a moving rectangle within a given observation window \(W\). In this paper, the distribution of the scan statistic is approximated using the formulae for the distribution of the maximum of a stationary sequence of \(1\)-dependent random variables. The approximation method involves calculation of the empirical distribution function of the scan statistics conditioned on the number of points of the Poisson process in \(W\). Finally, the precision of this approximation is compared numerically with the accuracy of other known methods.