On some methods to analyse
nonstationary nonlinear data.
Author: J.J.Zebrowski
Statistical and symbolic dynamics measures devised for the analysis of 24-hour human
heart rhythm which is inherently nonstationary are presented. The results are based on
over 5 years of research. Entropy-like measures, algorhithmic complexity and histograms of
symbolic words are used in conjunction with the surrogate data technique. It is shown that
statistical and sequential complexity may be separate properties of a given system. This
effect may be specific to bio-medical systems with their many fold and many level systems
of control. Examples of the application of the methods described to mathematical models
will also be given. Finally, the use of symbolic dynamics to analyse the interaction of
two systems is discussed.
|
