Readme file

SERIES B
Statistical Methodology

M. Jansen: Multiscale Poisson data smoothing
Journal of the Royal Statistical Society, Series B, volume 68 (2006)
, part 1, pages 27-48

The file hits20031205 contains the observations that are used as an illustration in the paper.

The data consist of 355 weekly observations of the total number of hits between Sunday morning 12.00 a.m. and Saturday night 11.59 p.m. on a web domain (i.e. a set of mutually linked web sites, some of which have been created during the experiment).

The series starts on Sunday, February 2nd, 1997, and ends on Saturday, November 22nd, 2003. This period of time saw the rise of popularity of the internet. As explained in the paper, some of the sudden changes in the weekly number of hits cannot be explained by the launch of new subsites on the web domain.

The algorithms that are used to generate the smoothed version of the observations can be downloaded as a subset of the Matlab routines called PiefLab, which is on the web site

http://www.cs.kuleuven.be/~maarten/software/pieflab.html

This web site has instructions on how to install the package.

Once this package has been installed, start Matlab, and type help PiefLab/Poisson and help PiefLab/Poisson/PoissonTests for more information on the routines used for this paper.

The software package also contains a Matlab routine, called hits20031205.m, that generates exactly the same data set.

For more information, please contact:

Maarten Jansen
Department of Mathematics and Computer Science
Technical University of Eindhoven
HG 9.25
PO Box 513
NL 5600 MB Eindhoven
The Netherlands

E-mail: mjansen@win.tue.nl

  • Dataset (hits20031205.m, size - 2kb)
Journals

SERIES A
Statistics in Society

SERIES B
Statistical Methodology

SERIES C
Applied Statistics

SERIES D
The Statistician