Scholarlink

<p>With the expansion of global beef trade, fraudulent practices in beef origin tracing have become increasingly frequent, severely disrupting market order and threatening food safety. This study aims to establish a high-precision origin tracing model based on mineral elements and stable isotopes combined with the XGBoost algorithm to distinguish beef from China and Argentina. Beef samples from China (<i>n</i> = 242) and Argentina (<i>n</i> = 235) were collected, and the ratios of four stable isotopes (δ<sup>13</sup>C, δ<sup>15</sup>N, δ<sup>2</sup>H, δ<sup>18</sup>O) and the contents of 52 mineral elements were determined using EA-IRMS, ICP-MS, and ICP-OES. The results showed significant differences between the two countriesʼ beef in 15 mineral elements or isotopes, including δ<sup>18</sup>O, Ni, Co, NIR, Sb, Cs, Cd, Sc, As, Nd, Ti, Ca, Cu, Mo, δ<sup>13</sup>C and δ<sup>2</sup>H. The constructed XGBoost classification model achieved an accuracy rate of 92.86% on the test set, outperforming traditional PLS-DA and random forest (RF). Feature importance analysis revealed that δ<sup>18</sup>O and Rh were the most critical indicators for distinguishing beef from the two regions. This study demonstrates the significant application potential of this technology combined with machine learning in cross-border beef origin tracing.</p> Graphical Abstract <p></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

Geographical Origin Discrimination of Beef from China and Argentina Using Multi-machine Learning Models

  • 期刊论文

  • 发表日期：

    2026年7月9日

查看全文
  • Xiaomeng Xiang,
  • Chaomin Zhao,
  • Liangzi Wang,
  • Tingyuan Zheng,
  • Runhe Zhang,
  • Niannian Chen,
  • Jing Zeng,
  • Chiying Wu,
  • Wenxuan Cai,
  • Xiaomin Zhu,
  • Xionghai Yi,
  • Dehua Guo,
  • Bing Liu

摘要

With the expansion of global beef trade, fraudulent practices in beef origin tracing have become increasingly frequent, severely disrupting market order and threatening food safety. This study aims to establish a high-precision origin tracing model based on mineral elements and stable isotopes combined with the XGBoost algorithm to distinguish beef from China and Argentina. Beef samples from China (n = 242) and Argentina (n = 235) were collected, and the ratios of four stable isotopes (δ13C, δ15N, δ2H, δ18O) and the contents of 52 mineral elements were determined using EA-IRMS, ICP-MS, and ICP-OES. The results showed significant differences between the two countriesʼ beef in 15 mineral elements or isotopes, including δ18O, Ni, Co, NIR, Sb, Cs, Cd, Sc, As, Nd, Ti, Ca, Cu, Mo, δ13C and δ2H. The constructed XGBoost classification model achieved an accuracy rate of 92.86% on the test set, outperforming traditional PLS-DA and random forest (RF). Feature importance analysis revealed that δ18O and Rh were the most critical indicators for distinguishing beef from the two regions. This study demonstrates the significant application potential of this technology combined with machine learning in cross-border beef origin tracing.

Graphical Abstract