Unveiling the drives behind tetracycline adsorption capacity with biochar through machine learning

Abstract

This study aimed to develop a robust predictive model for tetracycline (TC) adsorption onto biochar (BC) by employing machine learning techniques to investigate the underlying driving factors. Four machine learning algorithms, namely Random Forest (RF), Gradient Boosting Decision Tree (GBDT), eXtreme Gradient Boosting (XGBoost) and Artificial Neural Networks (ANN), were used to model the adsorption of TC on BC using the data from 295 adsorption experiments. The analysis revealed that the RF model had the highest predictive accuracy (R2 = 0.9625) compared to ANN (R2 = 0.9410), GBDT (R2 = 0.9152), and XGBoost (R2 = 0.9592) models. This study revealed that BC with a specific surface area (S (BET)) exceeding 380 cm3·g−1 and particle sizes ranging between 2.5 and 14.0 nm displayed the greatest efficiency in TC adsorption. The TC-to-BC ratio was identified as the most influential factor affecting adsorption efficiency, with a weight of 0.595. The concentration gradient between the adsorbate and adsorbent was demonstrated to be the principal driving force behind TC adsorption by BC. A predictive model was successfully developed to estimate the sorption performance of various types of BC for TC based on their properties, thereby facilitating the selection of appropriate BC for TC wastewater treatment.

Predictive capability of rough set machine learning in tetracycline adsorption using biochar

“…However, our earlier studies revealed that the initial TC concentration, BET surface area and pore volume had a significant positive correlation with the adsorption capacity of TC on BC (Zhang et al 2023). This is because e) molar ratio of hydrogen to carbon (H/C), (f) ash content (Ash, w%), (g) pH of the BC in water (pH_H 2 O), (h) BC pore size (PS, nm), (i) surface area (BET, m 2 .g −1 ), (j) pore volume (PV, cm 3 /g −1 ), (k) adsorption temperature (T, °C), (l) solution pH (pH_sol), (m) initial concentration of TC (mg/L −1 ), (n) initial concentration of BC (g/L −1 ), (o) the initial concentration ratio of TC to BC (C 0 , mmol/g −1 ) and (p) equilibrium sorption capacity Q e (mg/g.…”

“…To the best of the authors' knowledge, no prior literature exists that applies RSML algorithms to build performance prediction models for equilibrium sorption capacity for TC based on the combination of fifteen factors. However, few studies are available for predicting TC adsorption using BC through traditional ML methods (Zhang et al 2023;Zhou et al 2023). Thus, this study attempts to discover the applicability of rough sets in developing the prediction model that assists in choosing the optimized conditions for accomplishing maximum equilibrium sorption capacity.…”

Machine learning for persistent free radicals in biochar: dual prediction of contents and types using regression and classification models

“…In recent years, data-driven techniques, notably employing machine learning (ML) algorithms such as random forests (RF), support vector machines (SVM), and deep learning (DL) approaches like multilayer perceptron (MLP), have gained significant attention in constructing predictive models across various domains of environmental science (Zhong et al 2021), including contaminant monitoring (Gao et al 2021;Ullah et al 2023), micropollutant oxidation (Cha et al 2020), and new materials designing (Tang et al 2020). Recently, the scope of ML predictions has expanded to encompass diverse applications within biochar research, such as forecasting of micronutrients (Ullah et al 2023), heavy metal immobilization and migration (Li et al 2023), wastewater treatment (Kanthasamy et al 2023), antibiotic adsorption (Zhang et al 2023a), biochar functioning as biocatalyst (Wang et al 2023), and its impacts on GHGs emissions (Han et al 2024). These algorithms offer distinct advantages over traditional statistical methods by capturing nonlinear and complex relationships between features and target variables (Zhu et al 2020).…”