Yu Jiabin, Ma Xinyue, Zhang Xin, Cui Xiaoyu, Chen Shuaixiang, Zhao Zhiyao
School of Computer and Artificial Intelligence, Beijing Technology and Business University, Beijing, China.
NPJ Sci Food. 2025 Jul 9;9(1):132. doi: 10.1038/s41538-025-00495-8.
To solve the problem of subjectivity and low targeting of task-assigned food safety sampling, in this study, a targeted sampling decision-making method for food safety was proposed. First, a food decision-making factor reasoning module based on association analysis was constructed. An improved frequent pattern growth algorithm with constraints was used to mine food factor association rules based on decision-making factors. Second, a decision-making support module for targeted sampling to support food safety was constructed. A graph neural network was used to perform decision-making on the sampling frequency. The CRITIC-TOPSIS method was used to determine the sampling sequence of hazardous substances. In this study, experiments and analyses were conducted on sampling data of processed grain products in China (nationwide) and in Shandong Province from 2020 to 2022. Decision-making results of the sampling frequency and sampling order of hazardous substances were generated, thus demonstrating the wide applicability of the decision-making method.
为解决食品安全抽检任务分配主观性强、针对性低的问题，本研究提出了一种食品安全靶向抽检决策方法。首先，构建了基于关联分析的食品决策因素推理模块。采用改进的带约束频繁模式增长算法，基于决策因素挖掘食品因素关联规则。其次，构建了支持食品安全靶向抽检的决策支持模块。利用图神经网络对抽检频次进行决策，采用CRITIC-TOPSIS法确定有害物质的抽检顺序。本研究对2020—2022年我国（全国范围）及山东省粮食加工品的抽检数据进行了实验与分析，生成了有害物质抽检频次及抽检顺序的决策结果，从而验证了该决策方法具有广泛的适用性。
