论文部分内容阅读
应用径向基函数人工神经网络(RBF)和误差反向传播神经网络(BP),采用磺基水杨酸-邻菲罗啉双显色剂光度法同时分析了水中Fe(II)和Fe(III)。实验证明,RBF算法在所需神经元个数、训练时间、预测准确度等方面均明显优于BP网络算法。该技术和光度法结合有望成为多组分分析的有效选择方法之一。
The radial basis function artificial neural network (RBF) and error backpropagation neural network (BP) were used to analyze the effects of Fe (II) and Fe III). Experiments show that, RBF algorithm in the number of neurons required, training time, prediction accuracy are significantly better than the BP network algorithm. The combination of this technique and spectrophotometry is expected to be one of the most effective methods for multicomponent analysis.