Algorithms for interpretation of spectrometric data-a comparative study

The computer-based interpretation of spectrometric data usually involves the use of algorithms of deconvolution (or generalized deconvolution) for correction of two independent factors, the so-called spectral bandwidth of the instrument (SBW), and the natural bandwidth of the analyzed substance (NBW). In this paper, a comparative study of the interpretation efficiency of four algorithms- viz. the spectral deconvolution with the Tikhonov regularization, the iterative Jansson method, the Kalman filter with the positivity constraint, and the adaptive rational filter-is presented. The studied algorithms are compared with respect to the accuracy of the final results of interpretation, i.e. of the estimates of the positions and magnitudes of peaks of the absorption spectrum of the analyzed substance.