Statistical Computing and Statistical Graphics Newsletter, 1, 4-11, Oct. 1990

Feature Article:

Statistics for Chaos

Douglas Nychka, Stephen Ellner, Daniel McCaffrey and A. Ronald Gallant

North Carolina State University

Abstract

In the past twenty years much interest has been generated in the physical and biological sciences by deterministic mathematical systems that appear to have random, unpredictable behavior. These kind of phenomena, spanning a diverse range of fields, have been collected under the common heading of chaos. Models for chaotic behavior are important because they suggest a parsimonious representation for systems following seemingly complex behavior. This article will give a short introduction to chaotic systems and touch on some of the statistical issues in interpreting data gathered from a dynamic system. Since chaos theory provides a new model for observational data, it is a rich area for statistical applications. This topic is particularly appropriate for this newsletter because the analysis of such data inevitably requires intensive computing and graphics.