独立组分分析在红外光谱分析中的应用  

Application of Independent Component Analysis to the IR spectra Analysis

作　　者：毕贤[1] 李通化[1] 吴亮[1] 

机构地区：[1]同济大学化学系,上海200092

出　　处：《高等学校化学学报》2004年第6期1023-1027,共5页Chemical Journal of Chinese Universities

基　　金：国家自然科学基金 (批准号 :2 0 2 75 0 2 6)资助

摘　　要：针对红外光谱的黑色体系分析 ,提出了一种基于独立组分分析 ( ICA)的红外光谱定性分析方法 .其主要优点在于可从混合光谱中分离出独立组分的光谱 ,且这种分离是盲源分离 ,混合物的组分预先是未知的 .对 ICA在正己醇、丙酮和正丁醇的混合中红外光谱分离中的应用进行了研究 ,验证了体系的独立性 ,并对ICA算法做了一定的改进 ,讨论了各种预处理方法对其分离结果的误差 .A new method for separating the IR spectra based on the Independent Component Analysis(ICA) is presented. The main advantage of this method is that it can be used to separate the spectra of the independent components from the mixed spectra. The separation is a blind source separation,since the independent components in the mixtures was unknown. ICA was used to process the mixed system of \{n-hexyl\} alcohol,acetone and n-butyl alcohol in this study. We also discussed the effect of various preprocessing methods on the result of ICA and developed the ICA algorithm by adding PCA in front of ICA. [WT5HZ]

关 键 词：独立组分分析 ICA IR谱 应用 

分 类 号：O657[理学—分析化学]