Rapid detection method of the spicy components in Zanthoxylum bungeagum maxim by near infrared spectroscopy

Traditional method for detecting the spicy components in Zanthoxylum bungeagum Maxim was time-consuming and inefficient. To avoid these problems, a new approach based on near infrared spectroscopy (NIR) was developed. Firstly, 80 samples were selected from 141 powder samples with 40-mesh particle size by using principal component analysis and sample selection by genetic algorithms (PCA-SSGA). Then, the calibration model was established by near infrared spectra of 60 samples based on partial least squares (PLS), and applied to the test set with the other 20 samples. The results show that the determination coefficient (R-20(2)), the root mean square error of prediction (RMSEP20) the relative standard deviation ( RSD20) and the ratio of prediction to deviation (RPD20) are 0.9004, 0.8130, 6.07%, and 3.1670, respectively. Hence, the rapid detection of spicy components in Zanthoxylum bungeagum Maxim by NIR is feasible and efficient.