论文部分内容阅读
利用近红外光谱结合化学计量学方法快速检测了药用辅料糊精含量。首先,用近红外光谱仪采集主药与糊精共存样本的近红外光谱数据;然后,采用反向区间偏最小二乘法(Backward Interval Partial Least Squares,Bi PLS)优选光谱特征区间;最后,采用偏最小二乘法(PLS)对优选出来的区间建立药用辅料糊精近红外光谱模型。采用Bi PLS将全光谱均匀划分35个子区间,选择16个子区间[1,7,10,12,18,20,21,23~25,28~33]时,建立的模型预测效果最佳,其交叉验证均方根误差和预测均方根误差分别为1.501和2.437,校正集和验证集相关系数分别为0.9968和0.9958。因此,利用近红外光谱技术快速检测药用辅料糊精含量是可行的。
Near-infrared spectroscopy combined with chemometric methods for rapid detection of pharmaceutical excipients dextrin content. First, near-infrared spectroscopy was used to collect near-infrared spectral data of the co-existing sample of primary drug and dextrin. Then, the spectral interval was optimized by Backward Interval Partial Least Squares (Bi PLS). Finally, Second-order multiplication (PLS) establishes a near-infrared spectral model of the pharmaceutical excipient for the preferred interval. Using Bi PLS, the whole spectrum is evenly divided into 35 sub-intervals. When 16 sub-intervals are selected [1,7,10,12,18,20,21,23 ~ 25,28 ~ 33], the model predicts the best results. The root mean square error of cross validation and the root mean square error of prediction were 1.501 and 2.437 respectively, and the correlation coefficients of calibration set and validation set were 0.9968 and 0.9958, respectively. Therefore, the use of near infrared spectroscopy for rapid detection of pharmaceutical excipient dextrin content is feasible.