Determination of endpoint of procedure for processing radix rehmanniae based on analysis of UV spectral data of the water steam extracts by CWT and kernel ICA

The UV spectra of the water steam extracts (WSEs), not that of the conventional rehmanniae sample itself, during the procedure of processing radix rehmanniae were measured. The raw spectral data were preprocessed by continuous wavelet transform (CWT) for elimination of the noise and enhancement of the spectral resolution, and the independent components (ICs) were estimated by kernel ICA (KICA) from the CWT coefficient matrix. The results show the ICs that were estimated from rehmanniae samples from different producing areas are similar, and their change trends are also similar. The ICs' relative concentrations are gradually decreasing or increasing at the initial procedure 1 h to 14 h and then gradually tend to stable after 16 h. Based on the variation of the ICs' relative concentrations, the procedure endpoint can be determined at 15 h, while that of the conventional sensory analysis is 14 h to 20 h made by different operators. The proposed UV-CWT-KICA method for analysis of the UV spectra of the WSE provides an alternative relatively inexpensive approach for endpoint determination of the traditional Chinese medicine (TCM) processing procedure.