水生生物学报2026，Vol.50Issue(7)：62-69,8.DOI:10.3724/1000-3207.2026.2026.0037
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
PREDICTION MODEL FOR BIOACCUMULATION FACTOR OF PERFLUO-ROOCTANOIC ACID IN AQUATIC ANIMALS BASED ON RANDOM FOREST
摘要
Abstract
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.关键词
全氟辛酸/生物积累因子/随机森林/机器学习Key words
Perfluorooctanoic acid/Bioaccumulation factor/Random forest/Machine learning分类
资源环境引用本文复制引用
沈泓辰,杨方星..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[J].水生生物学报,2026,50(7):62-69,8.基金项目
国家重点研发计划(2024YFD2402203)资助[Supported by the National Key Research and Development Program of China(2024YFD2402203)] （2024YFD2402203）
