食品科学 ›› 2022, Vol. 43 ›› Issue (3): 121-128.doi: 10.7506/spkx1002-6630-20201224-277

• 营养卫生 • 上一篇    下一篇

基于小波分解-长短期记忆网络预测模型的酱卤肉制品安全预测分析

尹佳,陈翔,董曼,陈锂,郭鹏程,张涛,文红   

  1. (1.湖北省食品质量安全监督检验研究院,湖北省食品质量安全检测工程技术研究中心,湖北 武汉 430075;2.武汉理工大学计算机与人工智能学院,湖北 武汉 430070)
  • 出版日期:2022-02-15 发布日期:2022-03-08
  • 基金资助:
    “十三五”国家重点研发计划重点专项(2018YFC1603602)

Prediction and Analysis of Marinated Meat Product Safety Risk Using Wavelet Transform-Long Short-Term Memory Prediction Model

YIN Jia, CHEN Xiang, DONG Man, CHEN Li, GUO Pengcheng, ZHANG Tao, WEN Hong   

  1. (1. Hubei Provincial Engineering and Technology Research Center for Food Quality and Safety Test, Hubei Provincial Institute for Food Supervision and Test, Wuhan 430075, China; 2. School of Computer Science and Artificial Intelligence, Wuhan University of Technology, Wuhan 430070, China)
  • Online:2022-02-15 Published:2022-03-08

摘要: 为实现酱卤肉制品安全风险精准预警,本研究基于2014—2019年全国酱卤肉制品历史抽样检验数据信息,尝试将小波分解和长短期记忆网络(long short-term memory,LSTM)模型相结合,构建了全国31 个省份酱卤肉制品安全风险预测模型。结果表明,小波分解-LSTM预测模型对酱卤肉制品安全风险预测有较高的准确率,以湖北省为例,预测准确率为0.99,全国31 个省份的平均准确率为0.95,标准偏差为0.029,整体准确率较高,且准确率波动较小,说明建立的小波分解-LSTM模型可以适用于酱卤肉制品安全风险等级的精准预测,可为日常监管和食品安全风险预警提供技术支撑。

关键词: 酱卤肉制品;风险预测模型;小波分解;长短期记忆网络

Abstract: In order to achieve precise early warning of marinated meat product safety, this study attempts to construct a safety risk prediction model for marinated meat products from 31 provinces in China based on nationwide sample survey data from 2014 to 2019 using combination of wavelet transform (WT) and long short-term memory (LSTM). The results showed that the WT-LSTM model had a high prediction accuracy of 0.99 for samples from Hubei province and 0.95 for nationwide samples with a standard deviation of 0.029. The overall accuracy was high with small fluctuations. We concluded that the model can be applied to accurately predict the safety risk level of marinated meat products, and thus can provide technical support for daily supervision and food safety early warning.

Key words: marinated meat products; risk prediction model; wavelet transform; long short-term memory

中图分类号: