Improved partial least squares models for stability-indicating analysis of mebeverine and sulpiride mixtures in pharmaceutical preparation: A comparative study

Performance of partial least squares regression (PLSR) is enhanced in the presented work by three multivariate models, including weighted regression PLSR (Weighted-PLSR), genetic algorithm PLSR (GA-PLSR), and wavelet transform PLSR (WT-PLSR). The proposed models were applied for the stability-indicating analysis of mixtures of mebeverine hydrochloride (meb) and sulpiride (sul) in the presence of their reported impurities and degradation products. The work introduced in this paper aims to compare these different chemometric methods, showing the underlying algorithm for each and making a comparison of analysis results. For proper analysis, a 6-factor, 5-level experimental design was established resulting in a training set of 25 mixtures containing different ratios of the interfering species. A test set consisting of 5 mixtures was used to validate the prediction ability of the suggested models. Leave one out (LOO) and bootstrap were applied to predict number of PLS components. The GA-PLSR proposed method was successfully applied for the analysis of raw material (test set 101.03% ± 1.068, 101.47% ± 2.721 for meb and sul, respectively) and pharmaceutical tablets containing meb and sul mixtures (10.10% ± 0.566, 98.16% ± 1.081 for meb and sul).