Scholarlink

Flipped learning is an emerging teaching pedagogy gaining popularity in academia. In this methodology, students are given pre-recorded lectures to watch as homework before coming to the live classroom. In the classroom, students engage in activities like problem-solving, peer teaching, projects, and presentations. However, this methodology lacks a way to monitor student attention while watching the pre-recorded lecture videos, which could lead to a decrease in their learning outcomes. We address this problem by designing a multi-output deep learning model to categorize students as attentive or non-attentive. The proposed model is an autoencoder-classifier neural network that uses brain waves (EEG signals) collected from students while watching the lecture videos. The model analyzes the brain waves to classify whether a student was attentive or not while watching a particular lecture video. We conducted experiments using a dataset specifically collected for flipped learning pedagogy in the laboratory. Experimental results are evaluated with standard performance metrics such as precision, recall, F1-score, and accuracy. The proposed model outperforms other classification models, including EEGNet, Shallow ConvNet, Deep ConvNet, Siamese Neural Network (SNN), Support Vector Machine (SVM), Random Forest (RF), and K-Nearest Neighbor (KNN).

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

Classifying Students in Flipped Learning Pedagogy Exploiting EEG Signals and Deep Learning Techniques

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Arpit Sanghai,
  • Bidyut Kr. Patra

摘要

Flipped learning is an emerging teaching pedagogy gaining popularity in academia. In this methodology, students are given pre-recorded lectures to watch as homework before coming to the live classroom. In the classroom, students engage in activities like problem-solving, peer teaching, projects, and presentations. However, this methodology lacks a way to monitor student attention while watching the pre-recorded lecture videos, which could lead to a decrease in their learning outcomes. We address this problem by designing a multi-output deep learning model to categorize students as attentive or non-attentive. The proposed model is an autoencoder-classifier neural network that uses brain waves (EEG signals) collected from students while watching the lecture videos. The model analyzes the brain waves to classify whether a student was attentive or not while watching a particular lecture video. We conducted experiments using a dataset specifically collected for flipped learning pedagogy in the laboratory. Experimental results are evaluated with standard performance metrics such as precision, recall, F1-score, and accuracy. The proposed model outperforms other classification models, including EEGNet, Shallow ConvNet, Deep ConvNet, Siamese Neural Network (SNN), Support Vector Machine (SVM), Random Forest (RF), and K-Nearest Neighbor (KNN).