A Robust Spectrophotometric Method using Least Squares Support Vector Machine for Simultaneous Determination of Anti−Diabetic Drugs and Comparison with the Chromatographic Method

Abstract

In the present paper, the simultaneous spectrophotometric estimation of Metformin (MET) and Pioglitazone (PIO) in an antidiabetic drug called Actoplus MET based on least squares support vector machine (LS-SVM) was proposed. The optimum gamma (γ) and sigma (σ) parameters were found to be 825 and 90 with the root mean square error (RMSE) of 0.1343for MET, as well as 1000 and 350 with RMSE=0.4120 for PIO. Also, the mean recovery values of MET and PIO were 99.81% and 100.19%, respectively. Ultimately, the real sample was analyzed by High-Performance Liquid Chromatography (HPLC) reference method and the proposed procedure. Then, one-way analysis of variance (ANOVA) test at the 95 % confidence level was performed on achieved results from HPLC and LS-SVM methods. The statistical data of these methods showed that there were no significant differences between them.