A numerical comparison is made of algorithms for computing the dimension of attractors using the Grassberger-Procaccia correlation dimension and the Badii-Politi nearest neighbor approach. Using experimentally realizable data sets, the nearest neighbor method appears to have the better power law behavior when the attractor dimension is between 3 and 7.
%0 Journal Article
%1 kostelich1989
%A Kostelich, Eric J
%A Swinney, Harry L
%D 1989
%J Physica Scripta
%K attrctor considerations dimension practical
%N 3
%P 436
%T Practical considerations in estimating dimension from time series data
%U http://stacks.iop.org/1402-4896/40/i=3/a=030
%V 40
%X A numerical comparison is made of algorithms for computing the dimension of attractors using the Grassberger-Procaccia correlation dimension and the Badii-Politi nearest neighbor approach. Using experimentally realizable data sets, the nearest neighbor method appears to have the better power law behavior when the attractor dimension is between 3 and 7.
@article{kostelich1989,
abstract = {A numerical comparison is made of algorithms for computing the dimension of attractors using the Grassberger-Procaccia correlation dimension and the Badii-Politi nearest neighbor approach. Using experimentally realizable data sets, the nearest neighbor method appears to have the better power law behavior when the attractor dimension is between 3 and 7.},
added-at = {2013-11-22T11:28:09.000+0100},
author = {Kostelich, Eric J and Swinney, Harry L},
biburl = {https://www.bibsonomy.org/bibtex/215b03321fafe31762c635bf48684ad77/adisaurabh},
interhash = {c8e38f440e266f17c2f9b9c9cbc1bebf},
intrahash = {15b03321fafe31762c635bf48684ad77},
journal = {Physica Scripta},
keywords = {attrctor considerations dimension practical},
number = 3,
pages = 436,
timestamp = {2013-11-22T11:28:09.000+0100},
title = {Practical considerations in estimating dimension from time series data},
url = {http://stacks.iop.org/1402-4896/40/i=3/a=030},
volume = 40,
year = 1989
}