Hilbert-Huang transform is powerful method of nonlinear and non-stationary time series analysis. Two approaches, mirror periodic and extrema extending method, have been developed for handling the end effects of Hilbert-Huang transform. The mirror peri...
Hilbert-Huang transform is powerful method of nonlinear and non-stationary time series analysis. Two approaches, mirror periodic and extrema extending method, have been developed for handling the end effects of Hilbert-Huang transform. The mirror periodic method is based on the data distribution at ends. Two mirrors are placed at the extrema of both ends. The data within the mirrors are extended with mirror reflection to obtain a periodic sequence. With this method, the end effects of empirical mode decomposition and Hilbert transform are solved once forever. The mirror periodic method can get best results by placed the mirrors at the extrema, which have symmetric property of the data. Extrema extending method is implemented by adding two maxima and minima at the ends, which are defined by the two consecutive extrema of the data at ends. It is easy to perform and has equally well behalf as mirror periodic method. The extrema method is superior to the mirror periodic method when the data has strong asymmetric waveform.