Article
Keywords:
Poisson point process; asymptotically uniform distributions; weak convergence; variation distance; rate of convergence; Poisson hypothesis testing; distance method; quadrat count method; oscillating point patterns; isotropic Gaussian oscillations
Summary:
Oscillating point patterns are point processes derived from a locally finite set in a finite dimensional space by i.i.d. random oscillation of individual points. An upper and lower bound for the variation distance of the oscillating point pattern from the limit stationary Poisson process is established. As a consequence, the true order of the convergence rate in variation norm for the special case of isotropic Gaussian oscillations applied to the regular cubic net is found. To illustrate these theoretical results, simulated planar structures are compared with the Poisson point process by the quadrat count and distance methods.
References:
[4] P. J. Diggle:
Statistical Analysis of Point Processes. Academic Press, London, 1983.
MR 0743593
[7] K. Matthes J. Kerstan J. Mecke:
Infinitely Divisible Point Processes. J. Wiley & Sons, Chichester, 1978.
MR 0517931
[8] O. Persson: Distance methods. Studia Forestalia Suecica 15 (1964), 1-68.
[9] I. Saxl J. Rataj:
Distances of spherical contact in lattices of figures and lattice of figures with faults. In: Geometrical problems of image processing. Research in informatics. Vol. 4. (U. Eckhardt, A. Hübler, W. Nagel and G. Werner, ed.), Akademie-Verlag, Berlin, 1991, pp. 179-184.
MR 1111702
[10] C. Stein:
A bound for the error in the normal approximation to the distribution of a sum of dependent random variables. Proc. 6th Berkeley Symp. Math. Statist. Prob. 2 (1970), 583-602.
MR 0402873
[11] D. Stoyan W. S. Kendall J. Mecke:
Stochastic Geometry and Its Applications. Akademie-Verlag, Berlin, 1987.
MR 0879119