基于BP神经网络的土壤墒情预测精度研究--以肥东县为例

Research on Soil moisture Forecast Model based on BP Neural Network——a Case Study at Feidong County

本文以肥东县土壤墒情为例,选用BP神经网络模型,对土壤墒情预测进行探究。先介绍神经网络相关理论知识,然后针对模型建立需要确定的参数做了详细说明。着重比较了隔日和五日两种预测结果的精度,结果表明隔日预报精度要高于五日预测精度,得出结论 BP神经网络短期土壤墒情预报能取得较高的精度,对农业生产具有实际指导意义。

In the paper, taking soil moisture in Feidong County as an example, the BP neural network model was selected to explore soil moisture forecast. Firstly, the relevant theoretical knowledge of neural network was introduced. And then some of the model parameters was determined to do a detailed explanation. Based on the two prediction models, the results showed forecast accuracy for the next day was higher than that after five days. Overall BP neural network achieved a higher prediction accuracy of soil moisture forecast in the short term, which had practical significance for agricultural production.