二进小波神经网络用于示波计时电位测定

Application of dyadic Wavelet Neutral Network in Oscillographic Chronopotentiometric Determination

提出了二进小波神经网络的结构及算法 ,并用于单组分和多组分示波计时电位信号的浓度计算。在二进小波神经网络中选用了Morlet母小波和修正的误差反传前向神经网络。探讨了二进小波神经网络中小波基个数、初始学习速率因子和动量因子等参数对网络预测结果的影响。结果表明

The structure and algorithm of the dyadic wavelet neural network are described.The dyadic wavelet neural network is applied to predict quantitatively the concentration of mono\|component and double\|component in oscillographic chronopotentiometry.The network is constructed by the error back propagation neural network used Morlet mother wavelet basic function as node activation function.The effect of wavelet base number,learning rate factor and momentum factor on prediction is discussed.The experimental results show that this technique has higher convergence rate and prediction accuracy.