论文部分内容阅读
目的:应用可见/近红外光谱技术结合特征波长提取方法研究快速无损测定蜂胶毛胶中主要活性成分总黄酮含量的方法。方法:对112个蜂胶样本进行光谱扫描,并采用紫外分光光度法测得蜂胶总黄酮的化学值。经过平滑、变量标准化、一阶及二阶求导、去趋势等预处理,选择最优预处理方法。采用3种有效波长提取方法:连续投影算法(SPA)、载荷系数法(xLW)和回归系数法(RC),将提取的有效波长作为最小二乘-支持向量机(LS-SVM)、偏最小二乘法(PLS)和多元线性回归(MLR)的输入变量,分别建立了各种线性和非线性模型。结果:以预测集的预测相关系数(Rp),预测标准误差(RMSEP)和相对分析误差(RPD)作为评价指标比较各模型。结果表明,蜂胶毛胶中总黄酮含量预测的最优模型为SPA-PLS模型,Rp=0.954,RMSEP=9.32,RPD=4.02。结论:基于光谱技术和化学计量学方法的检测模型具有较好的检测效果,可作为蜂胶毛胶中有效成分的快速测定新方法。
OBJECTIVE: To study the rapid and non-destructive method for the determination of the total flavonoid content in the main active ingredient of propolis gummy gum by visible / near infrared spectroscopy combined with the characteristic wavelength extraction method. Methods: 112 propolis samples were scanned by spectrophotometry and the chemical value of total flavonoids was measured by UV spectrophotometry. After smoothing, variable standardization, first order and second order derivation, to trend and other pretreatment, choose the best pretreatment method. Three kinds of effective wavelength extraction methods were adopted: SPA, load factor (xLW) and regression coefficient (RC). The extracted effective wavelength was regarded as least squares support vector machine (LS-SVM) (PLS) and multiple linear regression (MLR), various linear and nonlinear models were established respectively. Results: The models were compared with the predictive set of prediction correlation coefficient (Rp), prediction standard error (RMSEP) and relative analysis error (RPD). The results showed that the best model for prediction of total flavonoids content in propolis was SPA-PLS model with Rp = 0.954, RMSEP = 9.32 and RPD = 4.02. Conclusion: The detection model based on spectroscopy and chemometrics method has a good detection effect and can be used as a rapid method for the determination of active ingredients in propolis Gums.