BULGARIAN CHEMICAL COMMUNICATIONS, cilt.49, sa.2, ss.410-421, 2017 (SCI-Expanded)
The primary aim of the present study was to compare the prediction power of different PLS algorithms as applied to the quantification of three spectrally overlapping drugs. Four variants of PLS were chosen for multivariate calibration and prediction of the three components of the drug formulation (paracetamol, propyphenazone and caffeine). NIPALS and SIMPLS algorithms were the most commonly used algorithms. The other tested algorithms were Kernel and Bidiagonalization which have been rarely applied in pharmaceutical analysis.