Prediction of storage tobacco mildew based on BP neural network optimized by beetle antennae search algorithm

摘要

For the improper storage, tobacco leaf mildew causes huge economic losses to the tobacco industry. It is particularly important to build a simple and accurate prediction model of tobacco leaf mildew. In this paper, BP neural network model is established, beetle antennae search algorithm is used to optimize the initial weights and thresholds of the BP neural network, a new BAS-BP model is proposed to predict the mildew of stored tobacco leaves. Simulation results show that BAS-BP model has better prediction accuracy than the standard BP neural network. Compared with PSO-BP model, the accuracy is close, but BAS-BP takes less time.