径向基神经网络示波计时电位法同时测定铅和镉

Simultaneous determination of Pb^(2+) and Cd^(2+) by oscillographic chronopotentiometry with radial basis function neural networks

将径向基神经网络用于Pb^(2+)和Cd^(2+)示波计时电位重叠切口的解析,建立了同时测定Pb^(2+)和Cd^(2+)的径向基神经网络示波计时电位分析新方法。实验表明:所建模型对训练集20个样本,预测结果相对标准偏差平均值仅为0．93%,而主成分回归模型预测结果为3．2%,偏最小二乘回归模型预测结果为3．0%。用本文方法预测的结果比主成分回归和偏最小二乘回归模型准确度高。

Radial basis function networks （RBFN） was applied to resolution overlapping oscillographic signals of Pb^2＋ and Cd^2＋ , and a new oscillographic chronopotentiometry with RBFN for the determination of Pb^2＋ and Cd^2＋ was presented. Experimental results showed that the relative standard deviations of prediction for calibration set obtained with RBFN, PLS and PCR methods were 0. 93%, 3.03%, 3.21%, respectively. This indicates that the RBFN is a valuable tool in multicomponent oscillographic analysis with the merits of simple operation and accurate prediction results.