Scholarlink

Educational institutions increasingly rely on data-driven approaches to enhance learning experiences, making student performance prediction a crucial aspect of learning analytics. This study applies deep learning models within the framework of Contextualized Learning Analytics to address this challenge. Using the Open University Learning Analytics Dataset (OULAD), the research explores the effectiveness of Long Short-Term Memory networks (LSTMs), Gated Recurrent Units (GRUs), Recurrent Convolutional Neural Networks (RCNNs), and Bidirectional LSTMs, with a focus on Recurrent Neural Networks (RNNs). Key components include rigorous data preprocessing, addressing class imbalances through data synthesis, and detailed dataset analysis. To evaluate the performance of the models, accuracy, precision, recall, and F1 score were used. The results provided valuable insights into the strengths and limitations of these models in adaptive learning contexts.

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

Contextualized Learning Analytics Using Deep Learning Models for Student Performance Prediction

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Maria Ahmed Minhas,
  • Akhtar Jamil,
  • Alaa Ali Hameed

摘要

Educational institutions increasingly rely on data-driven approaches to enhance learning experiences, making student performance prediction a crucial aspect of learning analytics. This study applies deep learning models within the framework of Contextualized Learning Analytics to address this challenge. Using the Open University Learning Analytics Dataset (OULAD), the research explores the effectiveness of Long Short-Term Memory networks (LSTMs), Gated Recurrent Units (GRUs), Recurrent Convolutional Neural Networks (RCNNs), and Bidirectional LSTMs, with a focus on Recurrent Neural Networks (RNNs). Key components include rigorous data preprocessing, addressing class imbalances through data synthesis, and detailed dataset analysis. To evaluate the performance of the models, accuracy, precision, recall, and F1 score were used. The results provided valuable insights into the strengths and limitations of these models in adaptive learning contexts.