Scholarlink

<p>Predicting student engagement is crucial for identifying the potential mental health challenges that may hinder academic performance and overall well-being. Early detection and intervention are essential to prevent detrimental effects on academic performance and overall quality of life. Effective strategies for predicting student engagement may involve analyzing various factors, this study employs a student engagement prediction model based on Improved LinkNet and Bidirectional Long Short-Term Memory (ImLN-Bi-LSTM), which considers face expression image and data features. Pre-processing, feature extraction, classification, and an engagement prediction mechanism are all included in the system. Face Expression image and data inputs perform individual pre-processing and feature extraction via distinctive approaches. The resultant features are then given to a hybrid classification model, utilizing Improved LinkNet and Bi-LSTM (Bidirectional Long Short-Term Memory) classifiers. The outcomes of ImLN-Bi-LSTM are the prediction results of student engagement in online learning. Comprehensive analyses including simulation and experimental assessments are conducted to validate the suggested ImLN-Bi-LSTM method. Moreover, at 80% of training data, the ImLN-Bi-LSTM model achieved a superior prediction accuracy of 0.951, and an F-measure of 0.905 which surpasses the result of traditional methods. The ImLN-Bi-LSTM model has the potential for use in online learning applications, and this study provides a solid and proven method for predicting student engagement.</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

Students’ engagement prediction in online learning context via face emotion and data features with improved LinkNet and Bi-LSTM architecture

  • 期刊论文

  • 发表日期：

    2025年8月4日

查看全文
  • Rama Bhadra Rao Maddu,
  • Murugappan S

摘要

Predicting student engagement is crucial for identifying the potential mental health challenges that may hinder academic performance and overall well-being. Early detection and intervention are essential to prevent detrimental effects on academic performance and overall quality of life. Effective strategies for predicting student engagement may involve analyzing various factors, this study employs a student engagement prediction model based on Improved LinkNet and Bidirectional Long Short-Term Memory (ImLN-Bi-LSTM), which considers face expression image and data features. Pre-processing, feature extraction, classification, and an engagement prediction mechanism are all included in the system. Face Expression image and data inputs perform individual pre-processing and feature extraction via distinctive approaches. The resultant features are then given to a hybrid classification model, utilizing Improved LinkNet and Bi-LSTM (Bidirectional Long Short-Term Memory) classifiers. The outcomes of ImLN-Bi-LSTM are the prediction results of student engagement in online learning. Comprehensive analyses including simulation and experimental assessments are conducted to validate the suggested ImLN-Bi-LSTM method. Moreover, at 80% of training data, the ImLN-Bi-LSTM model achieved a superior prediction accuracy of 0.951, and an F-measure of 0.905 which surpasses the result of traditional methods. The ImLN-Bi-LSTM model has the potential for use in online learning applications, and this study provides a solid and proven method for predicting student engagement.