首页 | 本学科首页   官方微博 | 高级检索  
     检索      

FA-RBF复合神经网络模型及其在地下水位预测中的应用
引用本文:黄华,刘玉邦.FA-RBF复合神经网络模型及其在地下水位预测中的应用[J].数学的实践与认识,2014(13).
作者姓名:黄华  刘玉邦
作者单位:重庆文理学院数学与财经学院;成都理工大学学术期刊编辑中心;
基金项目:国家自然科学基金(61271452);重庆市科委自然科学基金研究项目(2011jjA40029)
摘    要:地下水动态变化过程呈现出高度复杂的非线性特征,增加了地下水位预测的难度.为充分反映地下水位变化过程中自变量和因变量之间的非线性映射关系,克服在获取水文地质参数与查明水文地质条件方面的困难,避免部分智能方法实现繁琐复杂、计算效率低、限制条件多等不足,提出将因子分析方法与RBF神经网络算法构成复合模型,用于地下水位预测.结果表明,复合模型可以用于地下水位预测,模型计算结果可靠,网络训练时间缩短,计算精度有所提高;而且有成熟算法,实现简单.

关 键 词:地下水位预测  因子分析  径向基函数  神经网络

Research on FA-RBF Complex Neural Network Model and Its Application to Groundwater Level Prediction
Abstract:The groundwater dynamic process has been showing a high degree of complexity and nonlinear characteristics,which increases difficulty of forecasting the groundwater level.Currently,there are many methods of forecasting groundwater table which also have their own merits.However,some methods can not fully reflect the nonlinear relationship between the independent variables and the dependent variable,and some trapped in difficult to strike a hydrogeological parameters and to identify the hydrogeological conditions,and some intelligent methods are difficult to achieve due to more restrictive conditions and lower computational efficiency.Therefore,the factor analysis method and RBF neural network algorithm a composite model is used to forecast the groundwater level.A practical example shows that the composite model can be used to forecast the groundwater level completely.Composite model calculations reliable,and calculation accuracy is also improved.Model calculations have mature algorithm which is more easy to achieve.
Keywords:factor analysis  radial basis function  neural networks  groundwater level prediction
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号