Response Surface Methodology Validation of Zinc, Copper, and Lead Ions Adsorption Using Bayesian Regression

The adsorption of zinc, lead, and copper ions onto silica gel adsorbent has been successfully carried out in this study. Linear regression of polynomial transformation from input variables was employed to model the correlation between estimator variables (adsorbent dose, initial concentration, contact time, and pH) and output variable (%removal). Although the R² scores varied, overall, the models performed well in predicting metal ion removal. The regression coefficients of the models revealed that adsorbent dose and pH were the most significant factors for zinc and copper adsorption, while initial concentration and contact time also have a significant role in lead adsorption. Bayesian regression was used as a complementary approach to Response Surface Methodology (RSM), revealing different weight distributions for zinc and copper adsorption compared to RSM polynomial regression. The study concludes that copper and lead adsorption using RSM are more reliable compared to zinc, and suggests further optimization of factors or levels for more accurate results. The use of Bayesian regression provides valuable insights into variable weights and can improve the optimization process. Overall, this study provides useful information for designing efficient metal ion adsorption processes. This study provides useful insights for future research on the competition for metal ions in adsorption processes.