[A new approach to the fast measurement of content of amino acids in Cordyceps sinensis by ANN-NIR].

A new method for fast determining the content of amino acid in Cordyceps sinensis by means of near infrared (NIR) spectroscopy was developed. Colorimetry was first employed to measure the actual content of amino acid in Cordyceps sinensis. BP neural network was introduced to model the quantitative correlations between the NIR spectra and the contents of glycine, arginine and total amino acid. By comparing several preprocessing procedures and wavelength ranges, the optimal models could be obtained in the range of 7501.7-6097.8 cm(-1) and 5453.7-4246.5 cm(-1) with first derivative NIR spectra. Standard errors of prediction set (RMSEP) for glycine, arginine and total amino acid were 0.08, 0.07 and 0.36, respectively. The ultimate experimental results indicated that the proposed artificial neural network model was far superior to those of partial least square regression (PLS) and principal component regression (PCR). As an effective nonlinear multivariate calibration strategy, this paper could offer a new approach to the fast measurement of content of chemical components in traditional Chinese medicine by NIR spectroscopy.