迭代目标转换因子分析与人工神经网络法用于邻、间、对硝基甲苯的分光光度同时测定的比较研究  

Comparative Study on Simultaneous Determination of o -, m -, p -Nitromethylbenzene Using Spectrophotometry Combined With Iterative Target Transformation Factor Analysis and Artificial Neural Network Algorithm

作　　者：刘嘉[1] 邓勃[1] 

机构地区：[1]清华大学化学系,北京100084

出　　处：《分析化学》1995年第10期1172-1175,共4页Chinese Journal of Analytical Chemistry

基　　金：国家自然科学基金资助项目

摘　　要：本文将迭代目标转换因子分析与人工神经网络法用于分光光度法同时测定邻、间、对硝基甲苯,并与目标转换因子分析的结果进行了比较.结果表明,迭代目标转换因子分析法与线性网络法的效果都很好.其相对误差分别为1.3％和1.2％,而目标转换因子分析法的预测误差较大,其相对误差为10.4％.Iterative target transformation factor analysis (ITTFA) and artificial neural network (ANN) are used to determine o-, m-, p-nitromethylbenzene simultaneously with spec-trophotometry. After compared the results obtained from these two methods above mentioned with those from target transformation factor analysis, it shows that satisfied prediction precision could be obtained by ITTFA and ANN methods. The average relative errors of ITTFA and ANN are 1. 3% and 1. 2% respectively while target transformation factor analysis is 10. 4%. It also shows that the satisfied results can be obtained by artificial neural network when 3 layers network (L=3) and 10 neurons in each hidden layer(N=10) are adopted.

关 键 词：因子分析法 神经网络法 分光光度法 硝基甲苯 

分 类 号：O625.611[理学—有机化学] O657.32[理学—化学]