Feature Extraction for the Wrist-pulse-signals in Traditional Chinese Medicine by Ensemble Empirical Mode Decomposition

Pulse diagnosis plays a vital role in Traditional Chinese Medicine(TCM).To extract effective and properfeatures from the wrist-pulse-signals is a crucial step for the recognition and classification of the pulsesignals.Although the analysis in the time domain is also very effective for a lot of pulse types(SHU et al.,2007),the frequency and time-frequency analysis are usually used in the literature(YAN et al.,2005;YUEet al.,2006)because the pulse signals are non-stationary and non-periodic time series in the pathologicalstates.In recent years,a new time-frequency analysis method,Hilbert-HUANG Transform proposed by