我们旨在开发基于便携式傅里叶变换红外(FT-IR)光谱的关键质量特征预测算法(可溶性固体,水活动,pH值,蔗糖,葡萄糖,果糖,果糖/葡萄糖,羟甲基糠醛)的各种类型的糖蜜,确立其合法性,并创建一个模型来根据它们的植物起源将它们分开。标记为角豆树的样品(n=27),葡萄(n=24),Juniper(n=13),桑树(n=12)是从土耳其不同的当地市场购买的。在五个角豆树和七个葡萄糖蜜中发现了标签问题,通过参考分析证实了通过FT-IR算法分类为非真实的样品。为预测土耳其糖蜜的关键品质性状而生成的偏最小二乘回归模型显示出与参考分析(R2Val≥0.96)和低预测标准误差(SEP≤2.88)的良好相关性。FT-IR传感器为糖蜜测试提供了一种可行的方法,以通过制造和存储来评估其质量,还提供了一个强大的工具来确保正确的产品标签。
We aimed to develop portable Fourier transform infrared (FT-IR) spectroscopy-based prediction algorithms for the key quality characteristics (soluble solids, water activity, pH, sucrose, glucose, fructose, fructose/glucose, hydroxymethylfurfural) of various types of
molasses, establish their legitimacy, and create a model to separate them based on their botanical origin. Samples labeled as carob (n = 27), grape (n = 24), Juniper (n = 13), and mulberry (n = 12) were purchased from different local markets in Turkey. Labeling issues were revealed in five carob and seven grape
molasses, and those samples classified as non-authentic by the FT-IR algorithms were corroborated by reference analysis. Partial least squares regression models generated to predict the key quality traits of Turkish
molasses demonstrated excellent correlation with reference analysis (R2Val ≥ 0.96) and low standard error of prediction (SEP ≤ 2.88). The FT-IR sensor provided a feasible approach for
molasses testing to assess its quality through manufacturing and storage, also provided a powerful tool to -ensure proper product labeling.