Insights into the estimation of hydroxyl radical rate constant of water contaminants in AOP using new smart QSPR models

Abstract

AbstractInvestigations on the advanced oxidation processes (AOP) express that one of the effective oxidation agents in the degradation of organic pollutants is the aqueous hydroxyl radical. Therefore, the main aim of this work is the determination of the hydroxyl radical rate constant. To this end, three artificial intelligence methods were used to predict this constant for 457 various water pollutants from 27 molecular structures. The modelling results showed that the proposed decision tree model can determine the mentioned parameter with R‐squared values of 0.9264 and mean relative error of 1.479 for the overall data set. In addition, a comprehensive analysis of input parameters showed that Burden eigenvalue has the most impact on the hydroxyl radical rate constant. Finally, different comparisons were carried out between the suggested algorithms and other published approaches in the literature, which showed a high degree of precision for our models.