| With the advent of the era of big data,statistics and more and more disciplines have intersects.As the main content of modern statistics,statistical inference uses sample information to infer the overall distribution and digital characteristics,and mining the information behind the data.In recent years,food safety incidents emerge one after another,and the field of food safety has become a hot topic for statistical research.This paper focuses on the study of heavy metal concentration in aquatic products.The main contents are as follows:Firstly,the exploratory analysis of heavy metal data in aquatic products.In this study,the concentrations of five heavy metals(Pb,Hg,Cd,Cr,As)in four aquatic products(fish,shrimp,crab,and shellfish)from 32 provinces in China were analyzed during 2015-2017.By means of visualization methods such as box plot,map and cluster analysis,preliminary exploration of data was done.The distribution of sample points proved the validity of the data set,because its distribution pattern was consistent with the spatial distribution pattern of aquatic product production and consumption in China.The results of spatial distribution of heavy metal concentration indicated that the aquatic products with high heavy metal content were mainly concentrated in Qinghai(Cd,As),Xizang(Pb,Cr),Shanxi(Hg,Cd),Henan(Pb,Cr),Beijing(Cd)and Hainan(Hg).Clustering results showed that the contents of Pb and Cr in Tibet were much higher than those in other regions,which were 0.165 mg/kg and 0.916 mg/kg respectively.The Cr content in Macao and Guizhou was much higher than that in other regions,0.445 mg/kg and 0.603 mg/kg respectively,while the concentrations of other heavy metals were also high,so they were classified as one category.Secondly,the statistical inference of heavy metals in aquatic products.It mainly includes spatial autocorrelation analysis,non-parametric test and parameter estimation.The results of spatial autocorrelation analysis showed that the distribution of Pb and Cd in aquatic products was discrete and aggregated,while the distribution of Hg,Cr and As was random.For Pb,Yunnan was a "cold spot" area,that is,the concentration of Pb in this area presented low value.For Cd,Shaanxi was the "cold spot" region,while Zhejiang,Hebei,Tianjin and Beijing were the "hot spot" region.In other words,the content of Cd in aquatic products presented a low value aggregation in Shaanxi,while a high value aggregation in Zhejiang,Hebei,Beijing and Tianjin.The results of non-parametric test showed that there were significant differences in the concentrations of heavy metals in aquatic products from different regions,and the concentrations of heavy metals in North China,East China and South China were much higher than those in Southwest and Northwest China.The results of parameter estimation indicated that the concentration of heavy metals in crab and shellfish was higher than that in fish and shrimp.Maximum Likelihood Estimate and Expectation-Maximization algorithm had similar estimation results,but the variance of the latter estimation was smaller,the confidence interval was narrower,and the estimation accuracy was higher.Thirdly,the risk assessment of heavy metals in aquatic products.The non-carcinogenicity and carcinogenicity risk of heavy metals were evaluated by THQ/HI and TR coefficients.In addition,the results calculated by the average concentration and 97.5%quantile represent the risks caused by heavy metals to ordinary and high-consumption groups of aquatic products respectively.The THQ/HI calculation results showed that heavy metals in aquatic products in most areas did not pose a non-carcinogenic risk to human body,and high consumption groups of fish in North China should properly control the consumption of fish.TR calculation results suggested that high consumption groups of shrimp,crab and shellfish were easy to cause the carcinogenic risk of Cd to human body.High consumption groups of fish are easy to cause the carcinogenicity risk caused by Cr and As. |
PDF Full Text Request