基于主成分分析的土壤凋萎系数BP预测模型

发布时间：2018-06-26 12:44

本文选题：主成分分析 + 凋萎系数　；参考：《节水灌溉》2016年10期

【摘要】：基于黄土高原区农田耕作层土壤凋萎含水率的测试资料,建立了主成分分析与BP神经网络相结合土壤凋萎系数预测模型。通过主成分分析法减少了输入层神经元个数,优化了网络结构,提高了工作效率。预测值和实测值的相对误差平均值控制在5%以内,在可接受的范围,表明利用土壤基本理化参数预报农田耕作土壤的凋萎含水率是可行的。研究结果在提高传统神经网络的预测精度和收敛速度的同时,可为黄土高原区耕作农田作物用水管理以及促进土壤生产潜力的发挥提供强有力的理论支撑。
[Abstract]:Based on the test data of soil wilting moisture content in cropland of Loess Plateau, a prediction model of soil wilting coefficient combining principal component analysis (PCA) and BP neural network was established. The number of neurons in input layer is reduced by principal component analysis, the network structure is optimized and the working efficiency is improved. The average relative error between the predicted value and the measured value is controlled within 5%, which indicates that it is feasible to forecast the wilting moisture content of cultivated soil by using the basic physical and chemical parameters of soil. The results of the study not only improve the prediction accuracy and convergence rate of the traditional neural networks, but also provide a strong theoretical support for the management of crop water for farming in the Loess Plateau and for promoting the development of the potential of soil production.
【作者单位】：太原理工大学水利科学与工程学院;
【基金】：国家自然科学基金项目(40671081)
【分类号】：S152.7

【相似文献】