Abstract:

A three-layer BP network model was constructed with Qingyuan 9015 as the experimental material, and the mushroom growth days and the SO2-3 content in mushrooms, the SO2 content in fresh mushroom samples and the mushroom producing time (days) in picking period were used as the four input parameters, and the SO2 content in dried mushroom was used as output parameter. After 356 times of training process, the model converged with satisfying predictive ability.

Key words: artificial neural networks, mushroom, SO2