摘要
采用灰色系统理论中的关联分析方法,对影响土壤浸提液电导率的主要盐分离子类型进行分析,挑选出影响土壤浸提液电导率的主要因素。与此同时,介绍了径向基函数(RBF)神经网络的原理、训练算法,并建立了基于RBF神经网络与灰色关联分析的土壤浸提液电导率预测模型。结果表明,所建模型具有较强的非线性处理能力和逼近能力,运算速度快、性能稳定,预测精度较高,可以应用于生产实践中。
Using interaction analytical method in gray system theory, the main salty ion factors, which affect soil extract electrical conductivity, are analyzed and the main factors are selected. In the meanwhile, the principle and train algorithm of radial base function neural network are introduced and the prediction model of soil extract electrical conductivity based on RBF and grey relation analysis is established. The results show that the model has good nonlinear handling ability and approach ability, rapid operational speed, stable performance, and high forecast precision and can be applied to practice.
出处
《节水灌溉》
北大核心
2009年第7期15-17,20,共4页
Water Saving Irrigation
基金
国家高技术研究发展计划("863"计划)(2006AA100218)
新疆水利水电工程重点学科基金资助项目(XJZDXK2002-10-05)
关键词
灰色关联分析
电导率
RBF神经网络
grey relation analysis
electrical conductivity
RBF neural network
