基于自适应核独立成分分析的活性二氧化锰助剂光谱信息的提取

Extraction of Spectral Information of Additives from Activated Manganese Dioxide Products Using Adaptive Kernel Independent Component Analysis

利用乙醚、丙酮、氯仿、甲苯四种有机溶剂提取活性二氧化锰助剂,提取液经过滤浓缩后应用FT-IR(ATR附件)测定其红外光谱。采用绝对中位差法(MAD)确定混合体系中吸光物组分数,利用核独立成分分析(KICA)提取混合光谱信号中的纯组分光谱信息,建立了自适应核独立成分分析方法(AKICA)。采用AKICA提取得到的助剂光谱信息与所实际使用的化合物一致。结果表明AKICA具有无需经过繁冗化学或物理样品前处理而直接从混合物中直接提取光谱信息的能力,为混合体系中纯组分光谱信息的直接提取提供了新思路。

The additives were Abstracted from the manganese dioxide products with four kinds of organic solvents,ether,acetone,chloroform and toluene.The extracts were then baked and their attenuated total reflectance（ATR） FTIR spectra were measured using liquid membrane method.The number of chemical components of the additives was determined by median absolute deviation（MAD）,and the spectral information of the pure component was extracted by kernel independent component analysis（KICA）.The extracted spectral information of the additives is accordant to that of the practically used compounds.An adaptive kernel independent component analysis（AKICA） was proposed for directive extraction of spectral information from chemical mixtures.The results demonstrated that the AKICA method provides an alternative approach to extracting spectral information from the chemical mixtures without previously chemical or physical preseparation for direct extracting spectral information of pure components in the mixed system.