采用间隔偏最小二乘法(interval partial least squares,iPLS)、组合间隔偏最小二乘法(synergy interval partial least squares,SiPLS)、遗传偏最小二乘法(genetic algorithms partial least squares,GA-PLS)、竞争性自适应重加权法(competitive adaptive reweighted sampling,CARS)优选波长,并结合偏最小二乘法(partial least squares,PLS)建立白酒原酒中乙酸乙酯和乳酸乙酯定量分析模型。结果表明,上述4种方法都对模型有一定的优化效果,其中遗传算法结合组合间隔偏最小二乘算法(genetic algorithms-synergy interval partial least squares,GA-SiPLS)优选波长的优化效果最为明显,乙酸乙酯和乳酸乙酯的决定系数(R²)分别达到了0.989 7和0.991 0,预测均方根误差(root mean square error of prediction,RMSEP)分别为0.085 4、0.143 4,相对分析误差(relation percent deviation,RPD)分别为8.5和8.6,提高了模型的稳定性和精准性。说明近红外光谱分析技术对于白酒原酒中乙酸乙酯和乳酸乙酯含量的检测具有科学的指导意义。

In this paper, the contents of ethyl acetate and ethyl lactate in base liquor were quantitatively analyzed by near infrared spectroscopy. The characteristic wave bands were selected using interval partial least squares (iPLS), synergy interval partial least squares (SiPLS), genetic algorithm partial least squares (GA-PLS), competitive adaptive reweighted sampling (CARS), and partial least squares (PLS) to establish a model for quantitative analysis of ethyl acetate and ethyl lactate. The results showed that the above four methods had certain optimization effect on the model. The optimization effect of genetic algorithm partial least squares (GA-PLS) and synergy interval partial least squares (SiPLS) were the most obvious. The R² of ethyl acetate and ethyl lactate reached 0.989 7 and 0.991 0, and the Root Mean Squared Error of Prediction (RMSEP) respectively were 0.085 4 and 0.143 4, and the Ratio of performance to standard deviate (RPD) respectively were 8.5 and 8.6, which indicated that the stability and accuracy of the model were improved. The results showed that the near-infrared spectroscopy has a scientific significance for the detection of ethyl acetate and ethyl lactate in the base liquor.