Answers are grounded in this paper and cite their sources.
Reactive extraction is a promising separation technique for caproic acid recovery; however, evaluating the extraction efficiency (E%) often requires long experiments (16 h) and is also further complicated by physicochemical interactions. The present study explores the application of machine learning (ML) techniques to predict the E% for caproic acid systems using available variables prior to experimentation. Artificial Neural Networks, Random Forest Regressor, XGBoost, CatBoost, Gradient Boosting, kNN regressor, Huber Regressor, and Linear Regression were trained, evaluated, and employed for prediction. The XGBoost model shows the best performance, with the lowest mean absolute error of 1.44 by effectively capturing the non-linear relationships among acid concentration, solvents, and extractants. The interpretability of model outputs was assessed using Shapley Additive explanations and Permutation Feature Importance. The proposed ML framework reduces the experimental workload associated with E% prediction and supports process design through the integration of explainable AI.
No takes yet. Share an insight, caveat, or question.
Dahat et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: