地面水质评价的RBF神经网络方法  

RBF Network Method of Evaluating Water Quality

作　　者：李兴旺[1] 董曼玲[2] 

机构地区：[1]安徽水利水电职业技术学院,安徽合肥230601 [2]山东农业大学,山东泰安271018

出　　处：《水土保持通报》2002年第3期51-54,共4页Bulletin of Soil and Water Conservation

摘　　要：借助神经网络方法处理非线性问题的优势 ,采用径向基函数 (RBF)来构造多层前馈 BP神经网络。根据某流域水系的水质监测的数据 ,建立一个对地面水质进行判别的多层前馈网络数学模型。以地面水质污染主要的 7项指标为训练样本 ,利用该网络对水质进行评价 ,并将计算结果与其它方法进行比较分析。结果表明 ,该方法收敛速度较快 ,预测精度很高 。With the advantage of neural network in nonlinear problem, a radial basis function is used to improve conventional BP network. According to the condition of inspecting water quality in Dawen river of Huang river, the ANN model of evaluating water quality is put forward. The training stylebook takes 7 polluted target of this water quality as samples, and the water quality is evaluated using the trained network. The calculating results are analyzed and compared. This method is used to speed up the convergence and impove the performance.

关 键 词：地面水质 人工神经网络 BP网络 RBF网络 水质评价 

分 类 号：X824[环境科学与工程—环境工程]