小波神经网络用于钼和钨的同时测定

Simultaneous Determination of Molybdemum and Tungsten Using Wavelet Neural Network

介绍了小波神经网络的结构和算法，并首次用于混合试样中的钼和钨含量测定，在 小波神经网络中，采用Ｍｏｒｌｅｔ母小波和一维搜索变步长共轭梯度优化方法。结果表明，小波 神经网络优于ＢＰ网络，其预测含量的回收率在９６．０％～１０４．６％之间。

The wavelet neural networks have been presented and first applied to simultane- ous determination of molybdenum and tungsten. Morlet mother wavelet and line search con- jugate gradient optimization method are used in the networks. The results show that the wavelet neural networks are better than BP networks, and the predicted recoveries of molyb- denum and tungsten are between 96.0% and 104. 6%.