Rough set machine learning reveals governing factors of biochar-facilitated carbamazepine removal from water

Abstract

The presence of emerging pharmaceutical contaminants, particularly carbamazepine (CBZ), in wastewater has become a significant environmental concern due to its persistence in traditional treatment systems and potential adverse effects on aquatic ecosystems and human health. This study explores the efficacy of biochar, a carbon-rich material derived from biomass pyrolysis, as an adsorbent for removing CBZ from wastewater. A rough set machine learning (RSML) model was developed to predict CBZ removal efficiency. The model considered multiple operational parameters known to influence adsorption processes, including adsorption time, initial CBZ concentration, solution pH, adsorbent dosage, temperature, and adsorption type. The dataset was discretized to facilitate rough set analysis, allowing for identifying influential parameters and generating clear decision rules that link input conditions to removal efficiency. The results demonstrate that the RSML model attained a high classification accuracy of 93.15%, outperforming traditional classifiers. The model produced 49 scientifically coherent decision rules, providing valuable insights into the optimal conditions for maximising CBZ removal. This research highlights the potential of biochar as a sustainable solution for addressing pharmaceutical contaminants in wastewater and emphasises the importance of interpretable machine learning models in environmental engineering. The developed RSML tool offers practical guidance for real-time practitioners, enabling efficient and effective wastewater treatment strategies that can mitigate the ecological impacts of emerging contaminants like CBZ.