Scholarlink

<p>In the rapidly evolving landscape of online education, personalized learning experiences are crucial for enhancing student engagement and success. This research focuses on optimizing the e-learning journey by developing a hybrid recommender system that fuses Collaborative Filtering and Content-Based Filtering techniques. The innovative aspect of this approach lies in utilizing a Long Short-Term Memory neural network to dynamically determine the optimal weights for combining CF and CBF scores, thus ensuring that the system adapts to the evolving preferences of learners over time. The hybrid system leverages the strengths of CF, which utilizes historical user behaviour to suggest relevant courses, and CBF, which recommends content based on course attributes. The proposed system is implemented using python software. The LSTM network is trained on historical interaction data to capture temporal patterns and adjust the weights of CF and CBF in real-time, ensuring that recommendations remain relevant as user preferences change. Empirical evaluations demonstrate that the hybrid recommender system significantly outperforms standalone CF and CBF approaches in terms of precision, recall, and overall recommendation quality. The system’s adaptability, driven by the LSTM network, results in higher user satisfaction and engagement by continuously fine-tuning recommendations based on the latest user interactions. This study also highlights the system’s potential to offer valuable insights into learning patterns and trends, aiding educational institutions in curriculum design and content curation. Despite challenges such as the cold start problem, the hybrid approach shows promise for future enhancements, including the integration of user feedback and social interaction data. This research presents a robust solution for optimizing e-learning experiences through a hybrid recommender system that combines CF and CBF techniques with the adaptive capabilities of an LSTM network, paving the way for more personalized and effective online education.</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

Optimizing e-learning journey: fusing collaborative and content-based filtering in hybrid recommender systems

  • 期刊论文

  • 发表日期：

    2025年12月2日

查看全文
  • M. Ramesh,
  • V. R. Elangovan

摘要

In the rapidly evolving landscape of online education, personalized learning experiences are crucial for enhancing student engagement and success. This research focuses on optimizing the e-learning journey by developing a hybrid recommender system that fuses Collaborative Filtering and Content-Based Filtering techniques. The innovative aspect of this approach lies in utilizing a Long Short-Term Memory neural network to dynamically determine the optimal weights for combining CF and CBF scores, thus ensuring that the system adapts to the evolving preferences of learners over time. The hybrid system leverages the strengths of CF, which utilizes historical user behaviour to suggest relevant courses, and CBF, which recommends content based on course attributes. The proposed system is implemented using python software. The LSTM network is trained on historical interaction data to capture temporal patterns and adjust the weights of CF and CBF in real-time, ensuring that recommendations remain relevant as user preferences change. Empirical evaluations demonstrate that the hybrid recommender system significantly outperforms standalone CF and CBF approaches in terms of precision, recall, and overall recommendation quality. The system’s adaptability, driven by the LSTM network, results in higher user satisfaction and engagement by continuously fine-tuning recommendations based on the latest user interactions. This study also highlights the system’s potential to offer valuable insights into learning patterns and trends, aiding educational institutions in curriculum design and content curation. Despite challenges such as the cold start problem, the hybrid approach shows promise for future enhancements, including the integration of user feedback and social interaction data. This research presents a robust solution for optimizing e-learning experiences through a hybrid recommender system that combines CF and CBF techniques with the adaptive capabilities of an LSTM network, paving the way for more personalized and effective online education.