摘要
在全球化经济背景下研究中国出口欧盟食品的安全问题,食品安全受到多种复杂因素的影响,具有随机性、复杂性、非线性等特点,提出基于差分自动回归移动平均和支持向量机(ARIMA-SVM)的食品安全风险预测组合模型。利用ARIMA模型对食品出口时间序列进行建模,采用支持向量机对差分自动回归移动平均的预测残差进行建模,对两者结果进行权值相加以得到最终的食品安全预测结果,采用具体的食品安全数据进行仿真测试。结果表明,ARIMA-SVM组合模型具有更高预测精度和更低的平均绝对误差。
In the context of global economy,the food safety problems of China’s export to the European Union were studied.Food safety was affected by a variety of risk sources and had the characteristics of randomness,complexity and non-linearity.A combined model of food safety risk prediction based on differential automatic regression moving average and support vector machine(ARIMA+SVM)was proposed.The ARIMA model was used to model the time series of food export.The support vector machine was used to model the prediction residual of differential automatic regression moving average.The final food safety prediction was obtained by adding the weights of the two results.The results showed that the ARIMA-SVM model had higher prediction accuracy and lower mean absolute error.
作者
楼皓
曹倩
李海生
LOU Hao;CAO Qian;LI Haisheng(Beijing Key Laboratory of Big Data Technology for Food Safety,School of Computer and Information Engineering,Beijing Technology and Business University,Beijing 100048)
出处
《食品工业》
CAS
北大核心
2020年第1期334-339,共6页
The Food Industry
基金
国家自然科学基金青年基金(编号:61702018)
北京市属高校高水平教师队伍建设支持计划青年拔尖人才培育计划项目(CIT&TCD201804029).
