Answers are grounded in this paper and cite their sources.
Understanding the adsorption behavior of organic compounds is essential for advancing environmental and materials treatment processes. An interpretable quantitative structure–property relationship (QSPR) model based on multiple linear regression is developed to predict adsorption efficiency using simple descriptors derived directly from molecular formulas. Atom counts are combined with correction terms for π–π interactions and hydrogen bonding, preserving clear physicochemical meaning. Descriptor selection is performed by stepwise regression under strict criteria (P < 0.05, VIF<2) to ensure stability and robustness. The resulting equations show strong statistical performance (training r² = 0.977 and 0.948; external rₑₓₜ² = 0.937 and 0.947) with low prediction errors (RMSE = 0.148–0.252) and satisfactory external validation (CCCₑₓₜ = 0.956 and 0.970). Compared with descriptor-intensive models requiring specialized software, this approach remains simple, transparent, and easily reproducible while maintaining high accuracy. The novelty lies in representing complex adsorption phenomena through straightforward mathematical relationships based on molecular composition and mechanistically meaningful correction terms, providing a practical and interpretable framework for predicting adsorption efficiency.
No takes yet. Share an insight, caveat, or question.
Keshavarz et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: