[ Objective ]The research aimed to develop a new method for the detection of flavonoid compounds of Scutellaria bakalensis Georgi by near infrared reflectance spectroscopy (NIRS). [ Method ] Firstly, through the principal component analysis (PCA) of spectroscopic curves of 4 kinds of Scutellaria baicaknsis Georgi, the clustering of baicalin and wogonoside content of Scutellaria baiealensis Georgi was processed. Based on the PCA results,the first eight principal components were applied as artificial neural network (ANN) input nods, and 2 predictive indexes were applied as output nods, a testing model of near infrared reflectance spectroscopy with 8(input nods)-13(hidden layer nods)-2(output nods) was set up. [Result]The average relative error of NIRS model of baicalin and wogonoside content were 3. 87% and 5. 15%. The predicted value nearly was equal to HPLC value. The NIRS model had good predictability to analyze Scutellaria baiealensis Georgi quality. [ Conclusion ] NIRS model can be used on detecting Scutellaria baiealensis Georgi quality and quality controlling of Scutellaria baiealensis Georgi production processing.%[目的]提出一种用近红外光谱技术快速检测黄芩中黄酮类化合物的新方法.[方法]首先应用光谱仪获得4种黄芩的光谱曲线,用主成分分析法进行聚类分析,再结合人工神经网络技术建立模型进行检测.在主成分分析的基础上,以每一个样品的前8个主成分作为神经网络的输入节点,黄芩苷和汉黄芩苷2种成分类型作为神经网络的输出节点,建立一个8(输入节点)-13(隐含层节点)-2(输出节点)的3层人工神经网络模型.[结果]黄芩中黄芩苷和汉黄芩苷这2项指标人工神经网络模型预测值的平均相对误差分别为3.87%和5.15%,与高效液相色谱法测定值的符合程度很高,该模型具有很好的预测能力.[结论]新模型可用于黄芩的质量检测和生产加工过程中的质量控制.
展开▼
