傅里叶变换红外光谱结合化学计量学鉴别预包装纯菠萝汁中外源糖

摘要

This study was performed to establish a method that can quickly and accurately identify adulterated syrup in the pure pineapple juice .A attenuated total internal refraction‐fourier transform infrared spectroscopy was used to collect the range of 900~1 500 cm -1 infrared spectra of 234 samples pure pineapple juice and adulterated syrup by beet syrup ,rice syrup and cassava syrup .By using linear discriminant analysis and support vector machine for the identification model ,comparing the full spectral and selected wavelengths based on principal component analysis loading plots of the two models to identify adulteration .Studies showed that the correct rate of validation set by linear discriminant analysis and support vector machine model on full spectral were both higher than 88% ,variables were significantly reduced from 312 to 8 after selecting the eight characteristic wave‐lengths ,the correct rate of validation set by linear discriminant analysis model was up to 96.15% and support vector machine was increase to 94.87% .The results demonstrated that the model built using a attenuated total internal refraction‐fourier trans‐form infrared spectroscopy in combination with chemometric methods after selected characteristic wavelengths could be used for the identification of the adulterated syrup in the pure pineapple juice .%为了快速、准确鉴别预包装纯菠萝汁是否掺有外源性糖，采用傅里叶变换衰减全反射中红外光谱技术采集900～1500 cm－1范围内不同批次的预包装纯菠萝汁样品和掺入甜菜糖浆、大米糖浆、木薯糖浆的菠萝汁掺假样品的中红外光谱共计234例，以线性判别分析和支持向量机分析为掺假鉴别模型的建模方法，比较了全波段谱图与通过主成分载荷系数分析选取特征波长图谱的两种掺假鉴别模型。研究表明，全波段图谱的线性判别分析和支持向量机分类模型对验证集的判断正确率均高于88％；选取8个特征波长之后，线性判别分析模型验证集判断正确率提高至96.15％，支持向量机模型验证集判断正确率提高至94.87％，且模型输入变量由312个减少到8个。利用傅里叶变换衰减全反射中红外光谱技术结合化学计量学方法选取特征波长后建立的模型可以较好的应用于预包装纯菠萝汁外源糖的鉴别。