摘要
本文介绍人工神经网络及其在多组分红外分析中的应用。对于邻二氯苯、间二氯苯、对二甲苯和环己烷四组分红外分析,结果表明,当组分间相互作用或非线性性仪器响应因素引起的非线性响应存在时,人工神经网络能提供优良的光谱校准作用,因而为不经分离直接测定的多组分体系红外分析提供了一种新的途径。
This article describes the artificial neural networks and its application in the multi-component infrared analysis. The average recoveries of infrared determination of four-component o-dichlorobenzene , m-dichlorobenzene, p-dimethyl benzene and cyclobexane were 102. 3%, 100. 3% ,101. 6% ,and 100. 0% .respec- tively. This study indicates that when nonlinear response due to component interaction or due to nonlinear instrumental response functions are present , artificial neural networks may be capable of giving superior prefor-mance for spectroscopic calibration. Thus, the artificial neural networks may provide a new approach to determine the content from multi-component infrared spectroscopic data without any preliminary chemical separation.
出处
《分析化学》
SCIE
EI
CAS
CSCD
北大核心
1993年第4期435-438,共4页
Chinese Journal of Analytical Chemistry
关键词
人工神经网络
红外
分光光度法
Artificial neural networks , Multi-component infrared spectroscopic analysis ,o-Dichlorobenzene, m-Dichlorobenzene,p-Dimethyl benzene, Cyclohexane.
