The Journal of Light Scattering, Volume. 37, Issue 2, 299(2025)

Quantification of grape seed oil by fluorescence combined with Sammon mapping and support vector machine

SONG Su1,2、*, WANG Fangzheng2, GAO Jianan3, and LIU Hongsen4
Author Affiliations
  • 1School of Public Fundamentals, Jiangsu Vocational College of Medicine, Yancheng 224000, China
  • 2Department of Mathematics and Information Technology, The Education University of Hong Kong, Hongkong 999077, China
  • 3College of Artificial Intelligence, Nanjing Agricultural University, Nanjing 210031, China
  • 4School of Computer Science and Engineering, Southeast University, Nanjing 211189, China
  • show less

    Grape seed oil is a new type of vegetable oil that is rich in nutrients and has antioxidant and anti-inflammatory effects. However, the phenomenon of counterfeiting not only reduces the nutritional quality of grape seed oil but may also harm consumer health. To predict the concentration of contaminated grape seed oil, this paper proposes a detection method of adulterated grape seed oil using fluorescence spectroscopy technology combined with the combination algorithm of sammon mapping and support vector machine combination algorithm. Twenty-one kinds of simulated samples of adulterated grape seed oil were prepared, the fluorescence spectra of waste oil and grape seed oil were detected, and the spectral characteristics of waste oil and grape seed oil and their corresponding material information were analyzed. We extracted features from high-dimensional fluorescence spectral data using the sammon mapping algorithm. Then, we combined it with the support vector machine algorithm to predict the concentration of contaminated grape seed oil in gutter oil. Based on the sammon mapping and support vector machine combination algorithm of fluorescence spectra, we established a quantitative analysis model for detecting the concentration of contaminated grape seed oil in gutter oil. The determination coefficient of the quantitative model test set is 0.9560, and the root mean square error is 0.0040. The research results indicate a significant difference in the fluorescence spectra of gutter oil and grape seed oil, among which biological pigments are an important factor causing spectral differences. The combination algorithm of sammon mapping and support vector machine combination algorithm performs well in processing fluorescence spectrum data. Therefore, the detection method proposed in this study, which combines fluorescence spectroscopy technology with sammon mapping and support vector machine combination algorithm, has potential application value in predicting the adulteration concentration of grape seed oil. This study has a specific reference value for preventing and eliminating the adulteration of grape seed oil and provides methodological guidance for relevant departments.

    Tools

    Get Citation

    Copy Citation Text

    SONG Su, WANG Fangzheng, GAO Jianan, LIU Hongsen. Quantification of grape seed oil by fluorescence combined with Sammon mapping and support vector machine[J]. The Journal of Light Scattering, 2025, 37(2): 299

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: Sep. 25, 2024

    Accepted: Jul. 31, 2025

    Published Online: Jul. 31, 2025

    The Author Email: SONG Su (songsu1991@163.com)

    DOI:10.13883/j.issn1004-5929.202502018

    Topics