摘要
Abstract
In this study,we explored the combination of machine learning techniques and square wave anodic stripping voltammetry(SWASV)to improve the simultaneous detection of four heavy met-al ions:Cd2+,Pb2+,Cu2+,and Hg2+.Traditional electrochemical methods mainly rely on finding a linear response interval within a certain concentration range when detecting heavy metal ions,and in a multi-ionic environment,the SWASV curve often interferes,resulting in reduced accuracy.In this study,bare glassy carbon electrodes were used to detect repeatable SWASV of different concentra-tions of metal ion solutions,and important parameters such as current value,peak voltage and peak area were extracted from the detection data,and the concentration prediction model was constructed by combining extreme gradient boosting(XGBoost)and random forest(RF),and the support vector machine(SVM)was used.Among the machine learning classification algorithms,the SVR algorithm has the best effect(the area under the ROC curve of the four ions is greater than 0.95),and the fit degree(R-Squared)between the predicted value and the true value of the XGBoost concentration pre-diction model of the RF model is more than 0.95.By combining SWASV and machine learning,it is possible to achieve high-precision ion detection in complex ion mixing systems and improve the reli-ability of detection results.The results of this study provide an innovative solution for environmental monitoring and contamination control of multiple heavy metal ions,and demonstrate the application potential of machine learning in the field of electrochemical analysis.关键词
电化学/重金属离子/机器学习/干扰分析Key words
electrochemistry/heavy metal ions/machine learning/interference analysis分类
信息技术与安全科学
