Research On The Early Warning Method Of Food Safety Based On Neural Network And Quality Control Analysis

Food safety issues have become a major concern for the entire population.In recent years,most of China's major conferences have reported on the situation of food safety issues and made strategic arrangements.Zhang Mao,director of the Market Supervision Bureau,mentioned food safety supervision issues in the NPC and CPCPC conferences this year.Our country has realized the important impact of food safety on the development of the country and the happiness of the people,and will increase efforts to establish food safety risk standards.This paper constructs a dairy product safety evaluation index system.The reasonable warning level information can be provided and the appropriate recommendations can be given with the classification of risk levels.The main work of this paper is as follows:1)Establish an early warning model based on neural networks.The sample data is processed by the Interpretative Structural Model(ISM)and the Entropy Weight Based Analytic Hierarchy Process(AHP),which can obtain the first-level evaluation index of three types of dairy risk,and obtain the risk index of each level component.2)Perform data quality control analysis on the risk layer components.Taking the national limit standard for dairy products as the main reference,the obtained specification limits,control limits,capability index and performance index can help us to judge the discreteness of data and the quality of data security.Six early warning risk levels are formed by using the risk matrix.3)Set risk warning information through the construction process and data characteristics of the risk level.Each warning level corresponds to the specified risk component range,and comprehensively considers the impact of each component on the risk level,and gives early warning treatment suggestions.4)Develop and implement a food safety risk assessment and early warning prototype system,the results of which can be used to guide the specific detection range of quality control analysis.In order to verify the feasibility of the proposed risk warning model and data analysis method,this paper conducted several comparative experiments.The experimental results prove that the risk safety early warning evaluation results are reliable,and at the same time,it can effectively provide early warning decision support for government departments,including food safety departments.