FOOD SAFETY RISK PREDICTION METHOD BASED ON BRAIN NEURAL NETWORK

Food safety exists in various links such as food production, processing, transportation and sales, which affects the stable development of society. It has become an urgent problem for all countries in the world as how to make effective use of various safety inspection techniques, optimize the food processing and storage safety, predict potential food safety factors, and accurately assess and predict food safety risks. This study proposes a food safety risk prediction method based on brain neural network algorithm. Firstly, the key control points and index factors in the food safety supply chain are analyzed from the perspective of the food supply chain. Then the SOM self-organizing map and the K-means clustering method are used to select the data sets with high aggregation and low coupling to be used as training samples of neural network algorithm. Finally, three kinds of data are verified by BP neural network algorithm. The experimental results show that in food safety risk assessment and prediction, the data processed by two stages have better mean square error convergence, which increases the accuracy of neural network algorithm and improves the prediction effect. It provides a new prediction method for food safety risk prediction, which is of important practical significance.