Developed rapid spectrophotometric method for simultaneous quantitative determination of metformin and linagliptin mixture as antidiabetic drugs by artificial intelligence methodology in biological fluid and pharmaceutical sample

Abstract

In this study, simultaneous measurement of metformin (MET) and linagliptin (LIN) present in synthetic mixtures, antidiabetic tablet, and the biological sample was studied by chemometrics methods along with spectrophotometric technique. These procedures involved feedforward-backpropagation neural network (FFBP-NN) with Levenberg–Marquardt (LM) and comprehensive Grobner bases (CGB) algorithms, as well as least square support vector machine (LS-SVM). The mean square error (MSE) of the LM algorithm was obtained 1.22 × 10−28 (layer=2 with 10 neurons) and 1.93 × 10−27 (layer=3 with 8 neurons) for MET and LIN, respectively. In the CGB algorithm, layer 2 with 10 neurons for both components was introduced as the best layer and neurons with the least error. The optimal parameters of the LS-SVM method, including regularization (γ) and width (σ) parameters with the root mean square error (RMSE) of 0.4416 and 0.1634 were found for MET and LIN, respectively. The proposed methods were implemented on antidiabetic tablet and the results were compared with high-performance liquid chromatography (HPLC) using analysis of variance (ANOVA) test. The urine sample was also studied as a complex matrix by these methods. A combination of spectrophotometry and chemometrics can be used to improve the quality of pharmaceutical products due to its fast and low-cost and no need for the separation step.