基于BP神经网络算法的粮食储存品质预测模型构建研究

粮食安全预警对于储粮企业意义重大，本文基于BP神经网络算法，采用Python软件构建了粮食储存品质预测模型。通过使用全国各储粮生态区的质量品质数据来训练和测试，构建了基于BP神经网络算法的两种预测模型，时间序列品质预测模型可进行连续单指标跨期预测，温湿度-品质预测模型可进行度夏后粮情预测。经验证，时间序列品质预测模型对粮食储存品质指标预测效果较好，预测误差均低于15%。

Research of building grain storage quality prediction model based on BP neural network algorithm

Grain security early warning is of great significance to grain storage enterprises. Based on BP neural network algorithm, this paper uses Python software to build a grain storage quality prediction model.The prediction model based on BP neural network algorithm was trained and tested by using the quality data of the national grain storage ecological zones, and two application models were constructed: the quality prediction model of time series can be used for cross-period prediction on continuous single-index, the temperature and humidity - quality prediction model can be used to predict grain situation after summer. It has been proved that the quality prediction model of time series has a good prediction effect on grain storage quality index, and the prediction error is lower than 15%.