Multiresolution mode decomposition for adaptive time series analysis (Q2659750)

The author proposes a new model for adaptive series analysis called multiresolution mode decomposition based on a special form of a multiresolution intrinsic mode function to describe nonlinear and non-stationary data with time-dependent amplitudes, frequencies, and waveforms. For complex signals a recursive Gauss-Seidel iteration scheme is proposed to identify the multiresolution intrinsic mode functions and shape function series. The paper contains many numerical examples for various types of signals that confirm the efficiency of the proposed method.