Simultaneous multicomponent polycyclic aromatic hydrocarbon analysis using an independent component analysis-based latent variable regression

This article developed an independent component analysis-based latent variable regression (ICA-LVR) method, which is based on latent variable regression combined with independent component analysis. This strategy has been applied to the resolution of mixtures of four polycyclic aromatic hydrocarbons. Independent component analysis is a novel statistical signal processing technique based on the fourth-order moment of the signals aiming at solving related blind source separation (BSS) problem. Independent source variables and their corresponding concentration profiles can be extracted from the observed spectra of chemical mixtures. The independent source matrix instead of the original observed spectra combined with concentration matrix was used to build the regression model by latent variable regression (LVR). The method can obtain very selective information from unselective full-spectrum data. Experimental results showed the ICA-LVR method to be successful even where there was severe overlap of spectra and had the clear superiority over the LSV method.