PCA-SVR联用算法在近红外光谱分析烟草成分中的应用

Application of PCA-SVR to NIR Prediction Model for Tobacco Chemical Composition

由50份烟草样品的近红外漫反射光谱组成的光谱矩阵经过主成分分析降维,采用基于支持向量机回归(SVR)算法,以常规化学分析方法测定的总糖、还原糖、总氮、烟碱的含量为参考值,建立了烟草中主要成分近红外光谱定量分析定标模型,并采用留一法交叉验证(LOOCV)对模型进行验证。以内部交叉验证预测的RMSE值为判据,从核函数类型、惩罚因子C和不敏感函数ε取值等方面对定标模型进行优化,获得不同成分定标模型的优化参数。烟草总糖、还原糖、总氮、烟碱优化定标模型的RMSE值分别为1.581,1.412,0.117和0.313。同时建立了烟草以上成分的偏最小二乘回归(PLS)、多元线性回归(MLR)以及误差反向传播人工神经网络(BP-ANN)定标模型,通过内部交叉验证的RMSE值与SVR定标模型进行比较,结果表明SVR模型具有更好的预测效果。

Near infrared diffuse reflectance spectra of 50 tobacco samples were pretreated with PCA. The calibration models of determination of the main components in tobacco were developed with support vector regression （SVR）. The models were tested with leave-one-out （LOOCV） method and optimized with parameters of kernel function, penalty coefficient C and insensitive loss function. The root mean square errors （RMSE） with leave-one-out cross validation of the optimal models of nicotine, and total sugars, reductive sugar, and total nitrogen were 0. 313, 1. 581, 1. 412 and 0. 117 respectively. Based on the comparison of RMSE of the SVM model with those of the partial least square （PLS）, multiplicative linear regression （MLR） and back propaga- tion artificial neuron network （BP-ANN） models, it was found that the SVR model was the most robust one. This study suggested that it is feasible to rapidly determine the main components concentrations by near infrared spectroscopy method based on SVR.