Novel Iapso-Lstm Neural Network for Risk Analysis and Early Warning of Food Safety

After the food quantity is guaranteed, the food quality and safety are concerned. In order to improve the food quality and safety, we propose a novel adaptive particle swarm algorithm optimizing the long short-term memory neural network (IAPSO-LSTM) to make a risk early warning model of food safety in this paper. The risk value of the food safety detection data is obtained by the analytic hierarchy process using sum product and becomes the desired output of the LSTM. Then, the root mean square error is used as the objective function of the IAPSO algorithm, which is verified by five common benchmark functions, to find the optimal hyperparameter combination of the LSTM. Through the detection data being taken as the input of the IAPSO-LSTM, the risk early warning model of food safety based on the IAPSO-LSTM is established. Finally, the composite seasoning detection data in a Chinese province is taken as an example. The experimental results indicate that the overall performance of the IAPSO-LSTM is the best with the overall prediction error in term of the RMSE with the value being 0.474, the value of R 2 being 0.998 and the value of absolute error being less than 5. Moreover, the proposed risk early warning model can effectively help relevant government departments to warn of the potential risk factors in food and ensure the food safety.