Clustering data by inhomogeneous chaotic map lattices
Statistical Mechanics
2009-10-31 v2 Astrophysics
Chaotic Dynamics
Data Analysis, Statistics and Probability
Abstract
A new approach to clustering, based on the physical properties of inhomogeneous coupled chaotic maps, is presented. A chaotic map is assigned to each data-point and short range couplings are introduced. The stationary regime of the system corresponds to a macroscopic attractor independent of the initial conditions. The mutual information between couples of maps serves to partition the data set in clusters, without prior assumptions about the structure of the underlying distribution of the data. Experiments on simulated and real data sets show the effectiveness of the proposed algorithm.
Cite
@article{arxiv.cond-mat/0001445,
title = {Clustering data by inhomogeneous chaotic map lattices},
author = {L. Angelini and F. De Carlo and C. Marangi and M. Pellicoro and S. Stramaglia},
journal= {arXiv preprint arXiv:cond-mat/0001445},
year = {2009}
}
Comments
8 pages, 6 figures. Revised version accepted for publication on Physical Review Letters