Scholarlink

Abstract <p>The atmospheric pressure laser plasma ionization (<b>APLPI</b>) method, in combination with machine learning methods, is tested to solve the problem of vegetable oil classification. Samples of olive, rapeseed, sunflower, and linseed oils are studied. The samples are classified based on the mass-spectrometric profiles of volatile organic compounds emitted by the oils. It is shown that, in conducting hierarchical cluster analysis (<b>HCA</b>) with preliminary feature selection by the ANOVA method and reducing the dimensions of the response matrix using the <i>t</i>-distributed stochastic neighbor embedding (<Emphasis Type="BoldItalic">t</Emphasis><b>-SNE</b>), each type of oil forms a distinct cluster. Using an example of analyzing olive and rapeseed oil mixtures, it is demonstrated that a combination of the APLPI method with the multiple linear regression (<b>MLR</b>) method ensures the quantitative determination of the proportion of oils in the studied mixtures. The developed approach allows for the rapid and direct non-destructive analysis of vegetable oils without sample preparation and can be used to identify counterfeit products.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Direct Analysis of Vegetable Oils by Atmospheric Pressure Laser Plasma Ionization Combined with Machine Learning Methods

  • 期刊论文

  • 发表日期：

    2025年6月1日

查看全文
  • K. Yu. Kravets,
  • S. I. Timakova,
  • A. A. Grechnikov,
  • S. M. Nikiforov

摘要

Abstract

The atmospheric pressure laser plasma ionization (APLPI) method, in combination with machine learning methods, is tested to solve the problem of vegetable oil classification. Samples of olive, rapeseed, sunflower, and linseed oils are studied. The samples are classified based on the mass-spectrometric profiles of volatile organic compounds emitted by the oils. It is shown that, in conducting hierarchical cluster analysis (HCA) with preliminary feature selection by the ANOVA method and reducing the dimensions of the response matrix using the t-distributed stochastic neighbor embedding (t-SNE), each type of oil forms a distinct cluster. Using an example of analyzing olive and rapeseed oil mixtures, it is demonstrated that a combination of the APLPI method with the multiple linear regression (MLR) method ensures the quantitative determination of the proportion of oils in the studied mixtures. The developed approach allows for the rapid and direct non-destructive analysis of vegetable oils without sample preparation and can be used to identify counterfeit products.