LAUSR

Abstract As a type of amino acid salt, sodium glycinate (SG) is introduced as a potential solvent for CO 2 capture from gas streams like sour natural gas and flue gas. Modeling and assessment of the experimental data of CO 2 loading capacity of SG aqueous solution are the main objectives of the present study. To this end, a well-known ensemble method namely the Adaptive Boosting (AdaBoost) technique was utilized in conjunction with the Classification and Regression Tree (CART) algorithm for modeling the CO 2 solubility in SG solution as a function of the temperature of the system, CO 2 partial pressure, and concentration of SG in the aqueous phase. Then, the Leverage method was employed for investigating the range of applicability of the developed model and quality of the collected experimental data. Furthermore, the AdaBoost-CART was compared to multilayer artificial neural network (MLP-ANN) and least squares support vector machine (LSSVM). It was found that the presented AdaBoost-CART model provides more accurate predictions and has a wide range of applicability. Furthermore, the results suggest that among all the gathered data, six experimental equilibrium CO 2 solubility in SG solution are probable outliers.