Gaussian apodization evolving factor analysis as a novele method for curve resolution of gas chromatography–mass spectrometry datasets and precise local rank detection

چکیده مقاله

Gaussian-weighted moving window evolving factor analysis (GWMW-EFA) has been developed as a Gaussian apodization functions for factor analysis to assess the peak purity of the two-dimensional data. In GWMW-EFA method, rows (spectra) of submatrices which are extracted by fixed-size moving window, have been weighted by a Gaussian window, with the center of evaluated row (spectrum) and full width at half of the maximum (FWHM) equal to only one row width. Therefore, each submatrix mainly (89%) characterizes a single row and by performing factor analysis on this weighted submatrix, the number of principal components for each evaluated row, instead of all rows of a submatrix, is determined. Hence, the problem of time shifting in fixed-size moving window evolving factor analysis (FSMW-EFA) can be solved by this method. The results revealed that the GWMW-EFA algorithm could obtain proper initial estimation of elution profiles for curve resolution purposes in combination with alternating least-squares method.