Scholarlink

<p>In this work, Halim mediated Zinc oxide (H-ZnO) electrochemical nanosensor was used for copper detection in aqueous matrices and machine learning (ML) models are used for copper ion (Cu<sup>2+</sup>) classification. The nanosensor drop casted on a graphite electrode was employed to assess the sensing behavior using differential pulse voltammetry (DPV). The DPV experiment produced superior sensitive response for copper with a correlation efficient (R<sup>2</sup>) of 0.99. The quantified results were further used to build ML algorithms for classification. About 100&#xa0;DPV datasets in each were collected in the known concentration range of 0–1&#xa0;ppm. Feature extraction using t-distributed stochastic neighbor embedding (tsne) was preformed prior to classification. Copper classifications were performed using Decision Trees (DT), Artificial Neural Networks (ANN), k-Nearest Neighbor (kNN), and Support Vector Machines (SVM). Out of the classifiers, kNN was found to be outperforming with maximum accuracy values.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Machine Learning Approach for Copper Detection by Halim Mediated Zinc Oxide Nanosensor

  • 期刊论文

  • 发表日期：

    2025年10月28日

查看全文
  • K. Muthumeenakshi,
  • K. Vijayalakshmi,
  • S. Radha

摘要

In this work, Halim mediated Zinc oxide (H-ZnO) electrochemical nanosensor was used for copper detection in aqueous matrices and machine learning (ML) models are used for copper ion (Cu2+) classification. The nanosensor drop casted on a graphite electrode was employed to assess the sensing behavior using differential pulse voltammetry (DPV). The DPV experiment produced superior sensitive response for copper with a correlation efficient (R2) of 0.99. The quantified results were further used to build ML algorithms for classification. About 100 DPV datasets in each were collected in the known concentration range of 0–1 ppm. Feature extraction using t-distributed stochastic neighbor embedding (tsne) was preformed prior to classification. Copper classifications were performed using Decision Trees (DT), Artificial Neural Networks (ANN), k-Nearest Neighbor (kNN), and Support Vector Machines (SVM). Out of the classifiers, kNN was found to be outperforming with maximum accuracy values.