Summary
International Symposium on Nonlinear Theory and its Applications
2008
Session Number:A1L-D
Session:
Number:A1L-D4
Complementary Aspects of Spectral and Entropic Measures of Time-series
M.R. Titchener,
pp.-
Publication Date:2008/9/7
Online ISSN:2188-5079
DOI:10.34385/proc.42.A1L-D4
PDF download (1.9MB)
Summary:
In times past the term signal-analysis was synonymous with spectral analysis. There is however, a growing appreciation that most natural phenomena is non- linear, capable of exhibiting variable, indeed unpredictable behaviors not able to be expressed meaningfully in harmonic terms. A relative newcomer to an evolving range of non-linear tools is the T-entropy, a computable measure of information content for strings that may be applied to suitably encoded time-series. This paper highlights firstly the complementary aspects of the FFT and T-entropy in the characterisation of time-series but also their respective and contrasting sensitivities. The simple illustrations provided here of time-series comprising a mixture of harmonic and non-stationary components, subject to sampling and coarse graining, demonstrate effects which complicate interpretation of particularly the T-entropy results.