Rapid classification and quality assessment of citrus essential oils via machine-learning-assisted Raman spectroscopy

利用机器学习辅助拉曼光谱技术快速分类和评估柑橘精油的质量

阅读:4

Abstract

Citrus essential oils (EOs) require accurate identification and assessment to ensure authenticity and consistency. However, conventional techniques such as gas chromatography (GC) and mass spectrometry (MS) are time-consuming and expensive, highlighting the need for novel analytical methods. This study proposes an approach for EOs detection using Raman spectroscopy (RS) combined with machine learning (ML) algorithms. Six citrus EOs underwent an evaporation experiment, with Raman spectra collected at five time points and GC-MS was used to analyze compositional changes at the starting and ending points of evaporation as a standard reference. Five ML algorithms were developed to identify differences among EOs and monitor their temporal changes. The predictive performance was evaluated using multiple quantitative metrics. The results show that the support vector machine (SVM) consistently achieves the best performance across all prediction tasks, and the interpretability algorithm identified key components in the Raman spectra of EOs. Taken together, RS-SVM is proven to be an accurate and cost-effective analytical technique for the rapid identification and quality assessment of citrus EOs.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。