A Chemometric Analysis of Montelukast Sodium And Levocetirizine Dihydrochloride In Pharmaceutical Formulations Using Artificial Neural Networks, Partial Least Squares, And Principal Component Regression Models

Abstract

Present work describes applications and comparison of three models including Artificial intelligence-based ANN model and multivariate regression models namely PLS and PCR for simultaneous estimation of Montelukast sodium and Levocetirizine dihydrochloride in a tablet dosage form. Calibration and validation sets were prepared using standard solutions in a defined ratio as well as in random ratios. The UV absorption spectra of calibration set, and validation set were recorded in wavelength range of 200–400 nm using methanol as solvent. Three different models were constructed and validated using these sample sets. Three models were used to calculate the amount of drugs in the tablet formulation. Accuracy of the models was studied by performing recovery studies. Models were studied based on values obtained for root-mean-standard error for cross-validation, root-mean-standard error calibration, and correlation coefficients. All models showed lower values for the error terms and correlation values very close to one. Comparison of the three models shows that PLS, PCR and ANN models shows quite similar prediction ability, but ANN models need a larger number of samples to build a good model. These developed models were also compared with the reported HPLC method, and it was found that spectrophotometric based models are simple, rapid and gives better and faster prediction as compared to univariate models as well as costly, time consuming HPLC methods. Thus, these models may be used in the quality control department.