神经网络在四组分金属离子同时测定中的应用

非线性的方法解决了多组分金属离子体系同时测定的数据处理问题。该文研究了在不加缓冲液的条件下用EGTA直接络合滴定四组分金属离子时滴定体积V与溶液的pH值和溶液中金属离子的初始浓度礴之间的关系。在此我们用前向基多层神经网络(BP)和径向基神经网络(RBF)，发现在体系中RBF网络优于BP网络，结果令人满意。

The Application of Artificial Nerve Network to the Simultaneous Determination of Tetrad Metal Ions

Non - linearity method resolves the problem for the data of simultaneous measurement of the concentrations of mixtures of metal ions. This paper studies the complex ometric titration that we use, with ethy-lene glycol - bis(2 - aminoethylether) - N,N,N' ,N' - tetra acetic acid (EGTA) without buffer liquid, to get a relation among the titration volume V , the solution of pH and the original degree of ions solution CM . Here we use a multilayer feed- forward net with the backpropagation learning algorithm (BP) and a neural network based on radial basis functions, and discover that RBF network is better than BP network in the system.