Abstract: In order to improve the quality and yield of Agaricus bisporus and realize the early regulation of the overall environment of the mushroom house, it is the key to accurately predict the temperature and humidity data in the mushroom house environment. However, many parameters of the traditional prediction model need to be manually adjusted, such as the number of neurons in the hidden layer, the learning rate and the number of iterations. The selection of these parameters directly affects the prediction performance. In view of the above problems, a long short‑term memory (LSTM) mushroom house environment prediction model based on improved sparrow search algorithm (ISSA) optimization is proposed, which realizes the accurate prediction of the temperature and humidity environment in the mushroom house. The verification results show that the root mean square error and mean absolute error of the prediction method for temperature are 0.493 and 0.263, respectively, and the goodness of fit (R2) of the model is 96.2%. The root mean square error and mean absolute error of the humidity prediction index are 1.105 and 1.058, respectively, and the goodness of fit (R2) of the model is 95.6%. It can be seen that the proposed method is superior to the SSA—LSTM model in terms of the accuracy of temperature and humidity prediction in the mushroom house, and provides high‑efficiency decision data for creating the most suitable temperature and humidity environment in the mushroom house.

Key words: Agaricus bisporus room, Internet of Things, sparrow search algorithm, long short?term memory network

摘要： 为提升双孢蘑菇品质与产量，实现菇房整体环境提前调控，精准预测菇房环境中的温湿度数据是关键。但传统预测模型很多参数都需要人工手动调节，例如隐藏层神经元节点数、学习率、迭代次数等，这一系列参数的选择都直接影响预测性能的优劣。针对以上问题，提出一种基于改进麻雀搜索算法ISSA优化的长短期记忆网络LSTM菇房环境预测模型，实现对菇房内的温湿度环境的精准预测。验证结果表明：该预测方法对温度的预测指标均方根误差RSME、平均绝对误差MAE分别为0.493、0.263，模型拟合优度R2为96.2%；对湿度的预测指标均方根误差RSME、平均绝对误差MAE分别为1.105、1.058，模型拟合优度R2为95.6%，由此可见，在菇房温湿度预测准确率方面，所提方法优于SSA—LSTM模型，为打造最适宜的菇房温湿度环境提供高时效的决策数据。

关键词: 双孢蘑菇菇房, 物联网, 麻雀搜索算法, 长短期记忆网络