Artificial Neural Network-Genetic Algorithm based Prediction of Metformin Diffusion from Transdermal Films Based on Biopolymers for Diabetes Disease Treatment

The first aim of this study is to develop an intelligent model that combined the artificial neural network and the genetic algorithm (ANN-GA) to predict and simulate the Metformin (MFT) ex vitro skin permeation kinetics and the film mechanical properties. The second aim is to design matrix type transdermal films of MFT based on natural materials such as biopolymers which is a subject of great interest for the biomedical community. Chitosan (CTS) and Kappa Carrageenan (KC) in different ratios were used as matrix-forming agent using the solvent casting technique. Glycerin and Menthol were added to the films as a plasticizer and permeation enhancer, respectively. ANN-AG model accurately simulate the skin permeability of MFT from each formulation and predict the formulation parameters effect on the mechanical properties of the films. This performance was demonstrated by the obtained R2=0.9994 and RMSE=2.51x10-5. It has been observed that the increase in KC concentration has displayed increased value compared to the pure chitosan film. This may be attributed to the effect of polyelectrolyte complex (PEC) formation in CTS/KC films, especially at (1/1) ratio. The ex-vivo skin permeation study indicated high drug flux and good permeation enhancement effect.