Analysis and prediction of food safety risk based on sampling data of food supervision
Objective To understand the overall situation of random inspection of food safety in recent years,to play better in the role of random inspection of food supervision,and to enhance the targeting and pertinence of supervision.Methods Sampling inspection data of food supervision of Jizhou District Market Supervision Bureau of Tianjin from 2019 to 2021 were collected,and the factors that may lead to unqualified sampling inspection were analyzed by Chi-square test;Binary Logistic regression was used to establish a statistical model,and the factors that might affect the sampling results were explored from the perspective of food safety risk analysis,with the test level α=0.05.Results Rural compared with urban(x2=18.743,P＜0.05),catering compared with production and distribution(x2=32.606,P＜0.05),local compared with foreign(x2=23.349,P＜0.05),bulk compared with pre-packed(x2=32.542,P＜0.05).The unqualified rate of food sampling was high;Multiple Logistic regression analysis showed that sampling area,sampling year,food category and sampling season were the influencing factors of sampling results.The accuracy of prediction of sampling results by regression model was 78.42％,which could be applied in practice.Conclusions The binary logistic regression model can better predict the results of food sampling inspection.The supervision department should use the model to predict the results of food sampling inspection and identify the risk factors that lead to unqualified food sampling inspection,and supervision should be strengthened in rural areas,bulk food,local food,small and micro catering industry and so on,so that to prevent food safety risks and hidden dangers.
万方数据
维普
