基于随机森林的全氟辛酸水生动物生物积累因子预测模型OA

To quantitatively predict the bioaccumulation behavior of perfluorooctanoic acid(PFOA)in aquatic animals and provide a technical tool for the ecological risk assessment of emerging contaminants in aquatic environments and the safety management of aquatic products,this study developed a prediction model for the bioaccumulation factor of PFOA in aquatic animals using a Random Forest algorithm based on existing data.Waterborne PFOA concentration,water temperature,salinity,pH,dissolved oxygen,and species protein content were used as input variables,with BAF as the response variable,to characterize the nonlinear variation in bioaccumulation.The model demonstrated high predictive accuracy and strong generalization performance across both the training and testing datasets.Furthermore,the analysis revealed that waterborne PFOA concentration and organism protein content were the most influential factors contributing to the BAF of PFOA in aquatic animals.Overall,this study provides a convenient and reliable tool for quantitative predicting the bioaccumulation behavior of PFOA in aquatic animals and for supporting food-related health risk assessments based on the consumption of aquatic animals.