Enhancing the Spectral Analysis Efficiency at Low Signal-to-Noise Ratios Using the Technology of Surrogate Data without the Segmentation of Observation

摘要

A comparative analysis of two methods of estimating the observation correlation matrix (CM) based on obtaining an ensemble of its segments and a pseudo-ensemble obtained by applying the technology of surrogate data. It was shown that in the range of small values of the signal-to-noise ratio the error of CM estimation using the pseudo-ensemble was less than that of the CM estimation using the observation segmentation. Since the CM estimation is a basic procedure in advanced methods of spectral analysis, this study showed (by using simulation) that the application of observation pseudo-ensemble made it possible to avoid the segmentation of observations, which causes the reduction of spectral analysis resolution.