Comparison of abilities of two trend definition techniques for experimental data time series processing

[摘要] The abilities of two processing techniques for time series of physical data are studied. One of this techniques is based on the implementing of singular spectrum analysis, while the other is based on local approximation approach. The operation of these techniques is analyzed on a model time series. Both methods are compared in terms of accuracy of trend definition for different values of the noise level.

[发布机构] National Research Nuclear University MEPhI, Kashirskoe highway 31, Moscow; 115409, Russia^1;Geophysical Center of the Russian Academy of Sciences, Molodezhnaya St. 3, Moscow; 119296, Russia^2

[效力级别] 数学

[关键词] Local approximation;Noise levels;Physical data;Processing technique;Singular spectrum analysis;Time series processing