Research And Application Of Food Safety Risk Assessment Algorithm Based On Hidden Markov Model

"People rely on food as their top priority,and food as their priority".Food safety issues have a direct bearing on the people’s immediate interests,and are a major issue concerning the overall development of the economy,politics,and society.How to manage food safety issues has become a national strategy.In recent years,food safety incidents have occurred frequently at home and abroad.The poisoned egg incident in Europe in 2017 involved multiple countries,the salad poisoning incident occurred in the United States in 2018,and a major fish farming accident broke out in Norway in 2019.A series of food safety issues such as poisonous bean sprouts and Sudan red have also occurred in China.Food safety issues need to be resolved urgently,and scientific food safety risk assessment is of great significance.The current food safety risk assessment is a sampling inspection of food quality by the food inspection and inspection department.According to different food safety standards,the conclusions of qualified and unqualified products are obtained based on statistical methods.The unqualified food is mainly tested to reduce the risk.Food safety testing only evaluates according to the two statuses of "qualified" and "non-qualified",and it is impossible to dig deeper hidden risks of food safety information.Identifying the influential factors of food safety scientifically and accurately conducting risk assessment is vital to improving China’s food safety level.The research contents of this article are as follows:1.Aiming at the problem that traditional interpretive structural model(ISM)deals with the non-linearity and complexity of food safety data,an improved ISM method based on gray relational analysis(GRA)is proposed.GRA is used to calculate the correlation coefficient matrix between various influencing factors.Then the classification of influencing factors of food safety is realized and the main influencing factors are got,and it provides data support for determining index weight in food safety risk assessment.2.The GRA-ISM method is used to determine the weight of the detection index in combination with the influencing factors.A Hidden Markov Model(HMM)method based on the GRA-ISM method is proposed to evaluate the risk of food safety.The risk index is obtained by combining the corresponding weight of the index and the detection data.The 5-scale method is used to divide the risk into 5 levels as the observation sequence of the HMM model and then model the risk assessment.At the same time,the confusion matrix is used to analyze the accuracy and reliability of the model to verify the effectiveness of the proposed method.3.Aiming at the disadvantages of Baum-Welch algorithm used in HMM as a local optimal algorithm,it is easy to fall into the local optimal when training the model,and requires high initial parameters.This paper proposes an improved HMM method based on cuckoo search algorithm(CS).The CS algorithm is used to perform a global search to obtain the global optimal solution,and then the local optimal characteristics of Baum-Welch is combined with to make local corrections so that the parameters can quickly converge to the global optimal solution and improve the robustness of the algorithm.