Mathematical modeling of electrospinning process of silk fibroin/gelatin nanofibrous mat: Comparison of the accuracy of GMDH and RSM models

In this study, the nanofiber mat of silk fibroin and gelatin was fabricated by electrospinning process. As biological properties of the electrospun mat are bound up with the diameter of nanofibers, the group method of data handling along with the response surface methodology were adopted in order to explore the dependency of the diameter on the variables such as blend ratio, solution concentration, and electric field. The average absolute relative deviations were calculated as 4.62% and 5.86% for the group method of data handling and response surface methodology models, respectively. Moreover, being less prone to error and exhibiting greater accuracy to predict the actual trend of diameter in comparison with the response surface methodology, the group method of data handling model was applied to predict the porosity of silk fibroin/gelatin nanofiber mat. The average absolute relative deviations were 0.71% with a high regression coefficient of R ² = 0.99. This suggests that the group method of data handling model can fairly model the electrospinning process of the silk fibroin/gelatin nanofiber mat.