基于自助法和核密度估计的膳食暴露评估模型  

Food Safety Exposure Assessment Model Based on Bootstrapping and Kernel Density Estimation

作　　者：朱丰[1] 陆陶荣[1] 张立森[1] 

机构地区：[1]空军工程大学理学院,陕西西安710051

出　　处：《数学的实践与认识》2008年第14期59-68,共10页Mathematics in Practice and Theory

摘　　要：针对现阶段我国食品卫生安全保障体系的不完善,建立了基于自助法和核密度估计的膳食暴露评估模型.首先给出一种食物摄入量的抽样调查方案,并按污染物含量对特征人群进行模糊区间分类,建立特征人群每人每天食物摄入量的截尾正态分布模型.然后通过自助法重复抽样,弥补偶然抽查数据数量上的不足,提出了用对数正态分布核密度来估计污染物分布,并对模型进行了检验.在此基础上,建立了污染物摄入量模型,利用Gauss-Legendre求积公式求出其99.999%的右分位点,并且采用模糊匹配技术解决了数据不配套问题.最后对模型进行了讨论,给出了一种思路.The food safety exposure assessment model based on bootstrapping and kernel density estimation is established due to faultiness of food sanitation safety indemnification system at present in our nation. Firstly, a sample survey scheme of food consumption is given. The characteristic crowd is classified with the setting fuzzy interval according to contamination content. And truncated normal distribution model of daily food consumption for each person of characteristic crowd is established. Secondly, problem on lack of casual sample data is made up by means of bootstrapping repeat sampling. Logarithmic normal distribution kernel density is used to estimate contamination distribution, which is tested well. Based on these, contamination consumption model is established. And Gauss-Legendre quadrature formula is used to get 99. 999 % right-hand quantile. Moreover, the technology of fuzzy matching is used to solve inconsistent data. Finally, the model is discussed with another way of thinking.

关 键 词：食品安全 膳食暴露 自助法 核密度估计 

分 类 号：TS201.6[轻工技术与工程—食品科学]