Membrane adsorbents have emerged as effective engineering tools for the elimination of heavy metal ions from water resources. Our previous works presented novel modified chitosan/ poly(vinyl) alcohol membrane adsorbents for Cu(II) adsorption from water. This communication presents an expert mathematical model to predict the equilibrium adsorption of Cu(II) using Least-Squares Support-Vector-Machine (LS-SVM) intelligent approach. The model has been developed and tested using a total set of 72 experimentally measured equilibrium adsorption data. Membrane types, initial ion concentrations and temperature were selected as input parameters for the model. Equilibrium adsorption results were assigned as output parameters. The data points were categorized into three random sets: 70% assigned for the model development, 15% for validation and the remaining 15% for verification of the model applicability. The statistical Leverage analysis was employed for the examination of the outlier and doubtful data. According to the Williams plot and Hat matrix, all the data fall within the ranges of 0 6 H 6 0.055 and 1 6 R 6 1. These ranges are much better confidence limits relative to the requirement of the model. The results show that the developed model is quite accurate and reliable with the average absolute relative deviation of 4.8% and correlation coefficients close to unity. In addition, it is obtained that the model can appropriately predict the actual trend of the equilibrium adsorption as a function of initial metal concentration at different temperatures.
