基于BP神经网络的农业气象产量预报系统

Research on Weather Forecast System for Agricultural Yield Based on BP Neural Network

在深入研究农业气象产量预报系统和BP神经网络工作原理的基础上,针对现有系统预测精确性问题的不足,提出了基于BP神经网络的农业气象产量预报系统。在具体实现时,为了加快网络的训练速度,避免陷入局部极小和改善网络的概括能力,对BP算法进行了改进,利用北京地区的秋季水稻的产量为预测对象,通过网络数据的参数提取,运行验证系统达到了预测的目的,也为国家制定正确的农产品分配、供应、贮运与贸易等方面政策以及为计划,管理和安排等提供科学依据。

In thorough research on agricultural weather yield forecast system and BP neural network, a weather forecast for agricultural yield based on BP neural network was put forward to solve the shortage of the estimate-accurately problem. For the sake of the neural network training quickly, avoid getting into the part and improved the network generalize ability at designing, the BP algorithm was adapted. The autumn rice yield of Beijing region was made use for forecast object, the network running parameter was found.The system running result achieves the purpose of the forecast system, and its can provide guidance for country to adapt the policy for farm produce allotting, store and trade.