Support vector regression (SVR)-based adsorption model for Ni (II) ions removal

Abstract

Removal of heavy metals with adsorption using low-cost adsorbents has motivated many researchers across the world. In this work, support vector regression (SVR) model is developed for the prediction of the removal efficiency of Ni(II) ions using the waste from tea factory in terms of the independent parameters namely, particle size of adsorbent, pH of the solution, initial Ni (II) ion concentration, flow rate, effluent volume, bed depth and contact time. The SVR-based model is compared with the multiple linear regression (MLR) model using the statistical parameters. The correlation coefficient (R) and the average absolute relative error (AARE) values for the SVR-based has been obtained as 0.993 and 6.88% while those of MLR model they are 0.8393 and 74.54% respectively. The developed state-of-the-art SVR model, based on the structural risk minimization (SRM) principle, is found to predict the experimental values more accurately and is highly generalized.