Research On Risk Assessment Model Based On Recurrent Neural Network

With the improvement of social productivity and the development of social and economic level,people's living standards have been gradually improved,and people began to pay attention to the issue of food safety.The quality and safety of agricultural products has become the most concerned issue in today's society.Given that there are many risk factors affecting the quality of agricultural products in the links of growth and development,packaging and processing and logistics and distribution,a little carelessness in a certain link will lead to the quality decline of agricultural products.The establishment of the corresponding risk warning system,ensure the quality and safety of agricultural products,for the future improvement of agricultural products quality and safety risk warning system to provide a solid foundation,so as to effectively reduce food risks.To solve the above problems,the grasshopper optimization algorithm is adopted in this paper to optimize the key parameters in the gradient lifting decision tree algorithm,so as to improve the efficiency of the gradient lifting decision tree algorithm in feature extraction.The improved circular neural network is applied to the risk assessment model to improve the prediction accuracy of the model.The research contents of this paper are as follows:(1)The risk factors existing in the supply chain of agricultural products are analyzed in combination with the risk assessment principles,and the risk assessment index system is constructed.(2)In view of the risk characteristics in the risk assessment system of agricultural products,a group intelligence algorithm is proposed to optimize the key parameters in the GBDT algorithm,and the improved algorithm is used to extract the features and select the optimal feature subset.(3)Analysis of neural network model is applied to the feasibility and effectiveness of the risk assessment,in view of the traditional cycle length of neural network in the network of gradient disappeared,the cycle of the improved neural network algorithm is applied to the risk assessment,stretching the complexity of the network model,improve the nonlinear expression,enables the model to better predict risk.