Scholarlink

<p>Smart Class Internet of Things (IoT) integrates advanced internet-enabled devices to enhance learning environments through leveraging technologies like smart boards, IoT sensors, and student tracking systems. They allow for tracking of engagement in real-time, class attendance, and academic performance and enable interactive learning as well as management of classes. This study sought to establish the most appropriate Machine Learning (ML) model as well as the most appropriate attributes to apply to measure Smart Class IoT performance. ML schemes like Random Forest Regression (RFR) and Adaptive Neuro-Fuzzy Inference System (ANFIS) have been applied to process data from IoT. Optimization strategies like Improved Northern Goshawk Optimization (INGO) and Energy Valley Optimizer (EVO) have also been applied alongside the schemes to optimize the accuracy of prediction as well as the reliability of the systems. Feature selection was performed using the Partial Dependence Plot (PDP) method and retained only the most critical variables. Recursive Feature Elimination with Validation RFEV was more appropriate for prediction with <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="11760_2025_4707_Article_IEq1.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="20" /> </InlineMediaObject> <EquationSource Format="TEX">\(\:{R}^{2}\)</EquationSource> </InlineEquation> of 0.970, 0.985, and 0.750 for training, validation, and testing periods, respectively. Engagement Score was the most critical feature of them all, and it had a significant impact on model performance as evidenced through PDP analysis.</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

Optimization and feature analysis of smart class IoT systems for enhanced learning and classroom management performance

  • 期刊论文

  • 发表日期：

    2025年9月11日

查看全文
  • Junning Jiao

摘要

Smart Class Internet of Things (IoT) integrates advanced internet-enabled devices to enhance learning environments through leveraging technologies like smart boards, IoT sensors, and student tracking systems. They allow for tracking of engagement in real-time, class attendance, and academic performance and enable interactive learning as well as management of classes. This study sought to establish the most appropriate Machine Learning (ML) model as well as the most appropriate attributes to apply to measure Smart Class IoT performance. ML schemes like Random Forest Regression (RFR) and Adaptive Neuro-Fuzzy Inference System (ANFIS) have been applied to process data from IoT. Optimization strategies like Improved Northern Goshawk Optimization (INGO) and Energy Valley Optimizer (EVO) have also been applied alongside the schemes to optimize the accuracy of prediction as well as the reliability of the systems. Feature selection was performed using the Partial Dependence Plot (PDP) method and retained only the most critical variables. Recursive Feature Elimination with Validation RFEV was more appropriate for prediction with \(\:{R}^{2}\) of 0.970, 0.985, and 0.750 for training, validation, and testing periods, respectively. Engagement Score was the most critical feature of them all, and it had a significant impact on model performance as evidenced through PDP analysis.