engleski

A method of Noise Time-Series Characterization using Singular Value Decomposition

In this article we present an approach for characterization of noisy time-series originated from continuous chaotic processes. We used singular value decomposition (SVD) approach in reconstruction of the chaotic attractor from embedded vectors. We introduced a simple measure of embedded vectors directional distribution for chosen singular vector subspace, based on trajectory matrix nonlinear transformation. The value of such defined measure is dependent on amount of noise in the data. It is especially for data contaminated with isotropically distributed noise (or close to isotropic). That fact allows us to set up the criterion for minimal window width for attractor reconstruction as a function of noise in the data.

Chaotic attractor; Singular Value; SVD; Noise; Time series

nije evidentirano