An integrated model for predicting the size of silver nanoparticles in montmorillonite/chitosan bionanocomposites: A hybrid of data envelopment analysis and genetic programming approach

Unique chemical and physical properties of silver nanoparticles (AgNPs) enhances its usages in various categories such as medical utilities. Due to the high dependency of AgNPs properties to the size, this study is an attempt to employ gene expression programing (GEP) for constructing a quantitative model for estimating the size of AgNPs in montmorillonite/chitosan bionanocomposites that prepared by chemical approach. Generalization capabilities, fault tolerance, noise tolerance, high parallelism, nonlinearity and significant information processing characteristics are the main advantages of GEP. Accordingly, the practical parameters including reaction temperature, AgNO3 concentration, weight of montmorillonite in aqueous AgNO3/chitosan solution (WMMT) and the percentage of chitosan are selected as input parameters through GEP modeling. The accuracy of proposed models are investigated by statistical indicators including mean absolute percentage error (MAPE), root relative squared error (RRSE), root mean square error (RMSE) and correlation coefficient (R2). Finally, the best model is selected by R2 = 0.987, RMSE = 0.100, RRSE = 0.146 and MAPE = 0.221. The sensitivity analysis confirmed that the percentage of chitosan, concentration of AgNO3, WMMT and reaction temperature are the most effecting parameters on the size of AgNPs, respectively.