Comparison of LSSVM and RSM in simulating the removal of ciprofloxacin from aqueous solutions using magnetization of functionalized multi-walled carbon nanotubes: Process optimization using GA and RSM techniques.

This inquiry focuses on acquiring empirical models to predict ciprofloxacin removal using magnetization of functionalized multi-walled carbon nanotubes (FMWCNTs-Fe 3 O 4) from an aqueous solution. The response surface methodology (RSM) and support vector regression (SVR) as data mining techniques were adopted to develop models. Critical parameter effects comprising pH (3−10), adsorbent dose (0.2–1 g/L), contact time (5–60 min) and ciprofloxacin concentrations (30–100 mg/L) were analysed. The Langmuir, Freundlich, Temkin and Dubinin-Radushkevich isothermal models were utilized to fit the empirical data. FMWCNTs-Fe 3 O 4 prepared by chemical co-precipitation method was loaded by Fe 3 O 4 nanoparticles (using sonication) to synthesize functionalized multi-walled carbon nanotubes to remove ciprofloxacin (CIP). FMWCNTs-Fe 3 O 4 were characterized by fourier transform infrared spectroscopy (FTIR), X-Ray diffraction (XRD), scanning electron microscope (SEM), transmission electron microscope (TEM), vibrating sample magnetometer (VSM) methods. The Langmuir model was utilized to precisely describe the maximum adsorption capacity(q max) of 107.66 mg/g with R² = 0.998. In this study, the pseudo-second-order model exactly described the adsorption process(R² = 0.99). The results illustrated that the LSSVM (least squares support vector machine) model efficiently predicted the CIP removal percentage with very high accuracy in the training phase (R² = 0.975) and the test phase (R² =0.970). Moreover, the highest removal percentages in optimized step were achieved for RSM (pH 5.4, dose 0.78 g/L, time of 24.5 min, and CIP concentrations of 59 mg/L) and GA (genetic algorithm) (pH 4.4, dose 0.74 g/L, time of 42 min, and CIP concentrations of 38 mg/L) techniques by 88% and 99.1%, respectively. The FMCNTs-Fe 3 O 4 efficiency has decreased by 12% even after five used cycles relative to the optimal conditions (regeneration). Therefore, FMWCNTs-Fe 3 O 4 adsorption was considered to be an effective technique for CIP removal in the aqueous environment. [Display omitted] • RSM and LSSVM modeling to efficiency in adsorption of CIP from aqueous solutions using FMWCNTs-Fe 3 O 4. • The CIP removal percentage were accurately predicted by LSSWM model. • The Langmuir model was utilized to precisely describe the q max of 107.66 mg/g with R² = 0.998. • The removal percentages in the optimized step were achieved for RSM and GA techniques by 88% and 99.1%. • RSM and GA techniques' highest removal percentage in the optimized step 88% and 99.1% were obtained. [ABSTRACT FROM AUTHOR]