摘要
将人工神经网络(ANN)用于脉冲极谱法中Pb(Ⅱ)和T1(Ⅰ)重叠信号峰的解析,对神经网络参数的影响及优化作了研究。结果表明,网络的增益、学习速率和动量是影响网络收敛和稳定性的关键参数。本文还将神经网络与偏最小二乘法的计算结果作了比较。
An artificial neural network (ANN) was used to resolve overlapping pulse differ- ential voltammetric peaks of Pb(Ⅱ ) al1d Tl(Ⅰ ). The effects of network parameters were in- vestigated. Restults show that the gain, learning rate and momentum are critical for network convergence and stability. A comparison of ANN and partial least squares shows that these two methods are comparable.
出处
《分析化学》
SCIE
EI
CAS
CSCD
北大核心
1997年第3期249-252,共4页
Chinese Journal of Analytical Chemistry
基金
国家教委留学回国人员科研启动基金
关键词
人工神经网络
脉冲极谱法
ANN
极谱
Artificial neural network, chemometrics, partial least squares, pulse differential voltammetry