Resumo

PENAFIEL, V. M.. NONLINEAR TESTS BY SURROGATE DATA ON EXPERIMENTAL SERIES. Revista Boliviana de Física [online]. 2013, vol.22, n.22, pp.8-13. ISSN 1562-3823.

The method of surrogate data is introduced to test the nonlinear character of experimental time series, in particular, as a previous stage to test chaotic behaviour. We used several algorithms specifically oriented towards this purpose and applied to geomagnetic records (components D, H, Z and 3 time series with 2048 points each and taken at intervals of 1 minute) in order to test their discrimination capacity while studying series that could be interesting (because of their nonlinear character). We apply Fourier and simulated annealing methods to construct surrogate data. For the latter, and restricted by the invariance of the autocorrelation function, the sampling of the coeficients of the nonlinear prediction shows a remarkable discrimination capacity as compared to other methods.

Palavras-chave : Time series analysis; nonlinear dynamics; chaos.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )