A new approach in adsorption modeling using random forest regression, Bayesian multiple linear regression, and...

Area of Science:

• Environmental Chemistry
• Water Treatment Technologies
• Adsorption Science

Background:

• Mathematical modeling is increasingly vital for water quality prediction.
• Rice husk biochar shows potential for removing pollutants like 2,4-D from wastewater.
• Understanding adsorption parameters is key to optimizing water treatment processes.

Purpose of the Study:

• To evaluate Random Forest Regression (RFR), Bayesian Multiple Linear Regression (BMLR), and Multiple Linear Regression (MLR) for predicting 2,4-D removal.
• To identify key operating parameters influencing 2,4-D adsorption by rice husk biochar.
• To assess the efficiency of different regression models in a water quality monitoring context.

Main Methods:

• Utilized RFR, BMLR, and MLR models to predict 2,4-D elimination.
• Input parameters included initial 2,4-D concentration, adsorbent dosage, pH, reaction time, and temperature.
• Adsorption equilibrium, kinetics, and thermodynamics were analyzed using Freundlich, pseudo-first-order, and thermodynamic models.

Main Results:

• RFR achieved the highest prediction accuracy (R²=0.994, RMSE=1.92) compared to BMLR and MLR.
• Adsorption followed Freundlich and pseudo-first-order kinetics, indicating exothermic and spontaneous processes.
• Initial 2,4-D concentration and adsorbent dosage were identified as the most sensitive parameters.

Conclusions:

• RFR demonstrates superior performance in predicting 2,4-D removal due to its ability to handle non-linear relationships.
• The developed models can facilitate cost-effective and permanent water quality monitoring.
• Rice husk biochar is a promising adsorbent for 2,4-D removal, with optimized parameters enhancing efficiency.