Fast Determination of Active Components in Scutellaria Extract by NIRS Combined with PLS

In this paper, the three active components in Scutellaria Extract were rapidly determinate by near-infrared reflectance spectroscopy (NIRS). 100 samples of Scutellaria Extract from 12 different pharmaceutical factories were divided into a calibration set (80 samples) and a validation set (20 samples). In combination with the practical least square (PLS), the quantitative calibration models were established, involving baicalin, baicalein and wogonin. The correlation coefficient of cross-validation (R²_cv) was 0.99481, 0.99867 and 0.99484 for baicalin, baicalein and wogonin, the root-mean-square error of calibration (RMSEC) was 0.440, 0.0225 and 0.0111, the root-mean-square error of cross-validation (RMSECV) was 2.259, 0.0553 and 0.0483. The validation samples were used to evaluate the performance of the models, the correlation coefficient of prediction (r²) was 0.9882, 0.9965 and 0.9909, the root-mean-square error of prediction (RMSEP) was 0.486, 0.0271 and 0.0110. The results indicated that NIRS can provide a simple and accurate way in the determination of the baicalin, baicalein and wogonin content in Scutellaria Extract.