An Improved BP Neural Network Algorithm for Evaluating Food Traceability System Performance

BP neural network algorithm has powerful calculation ability, but the algorithm has some shortages such as low convergence which limits its application, so improving BP algorithm has become a matter of concern in the fields related. Based on analyzing improvement methods wildly used today, the paper presents a new BP neural network algorithm and applies it to evaluate food traceability system performance. Firstly, the paper improves the BP algorithm through changing learning rate, trigonometric function to simplify the original calculation structure; secondly, the calculation step of the improved BP algorithm is redesigned to speed up its convergence. Finally, the paper conducts the theoretical analysis of the calculation performance of the improved algorithm and applies it to evaluate food traceability system performance, the theoretical analysis and experimental evaluation results show that the improved algorithm can improve evaluation accuracy and algorithm calculation efficiency and can be used for evaluating food traceability system performance practically.