A hybrid model of support vector regression with genetic algorithm for forecasting adsorption of malachite green onto multi-walled carbon nanotubes: central composite design optimization

Abstract

The aim of this work is the study of the predictive ability of a hybrid model of support vector regression with genetic algorithm optimization (GA–SVR) for the adsorption of malachite green (MG) onto multi-walled carbon nanotubes (MWCNTs). Various factors were investigated by central composite design and optimum conditions was set as: pH 8, 0.018 g MWCNTs, 8 mg L⁻¹ dye mixed with 50 mL solution thoroughly for 10 min. The Langmuir, Freundlich, Temkin and D–R isothermal models are applied to fitting the experimental data, and the data was well explained by the Langmuir model with a maximum adsorption capacity of 62.11–80.64 mg g⁻¹ in a short time at 25 °C. Kinetic studies at various adsorbent dosages and the initial MG concentration show that maximum MG removal was achieved within 10 min of the start of every experiment under most conditions. The adsorption obeys the pseudo-second-order rate equation in addition to the intraparticle diffusion model. The optimal parameters (C of 0.2509, σ² of 0.1288 and ε of 0.2018) for the SVR model were obtained based on the GA. For the testing data set, MSE values of 0.0034 and the coefficient of determination (R²) values of 0.9195 were achieved.