Scholarlink

<p>This paper presents a hybrid intelligent course recommendation system that delivers customized educational routes to students. The system solves challenges of effective course recommendations through employing various recommendation approaches, such that the drawbacks of one method are covered up for by another. The recommendation system implements three recommendation strategies through content-based filtering for description and material analysis as well as collaborative filtering for user collective preferences followed by knowledge-based filtering for structured course prerequisites and learning objectives. We build transparency in the system along with user trust using Large Language Model (LLM)-based explanations which provide justification details for the recommended items. The system responds to the essential requirement of explainable intelligent systems. Our system receives evaluation through an extensive dataset that includes user behaviour records with additional information about courses and professional objectives. The system proves its ability to create appropriate recommendations while ensuring diversity and personalization thus solving the typical challenges of course discovery for new subjects and recommendation uniformity. The F1-score of the model comes out to be around 0.8, which is higher than the average ratings for traditional systems, which usually lie in the range of 0.57–0.72. These advances represent critical milestones in developing smart educational platforms which provide individualized support for lifelong learning journeys.</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

Augmenting Course Recommendations with Explainable AI: A Hybrid Approach Utilizing Large Language Models

  • 期刊论文

  • 发表日期：

    2025年7月11日

查看全文
  • Ahan Ganguly,
  • Ankan Dey,
  • Sourik Bhuiya,
  • Indranil Roy Banik,
  • Subhabrata Sengupta,
  • Rupayan Das

摘要

This paper presents a hybrid intelligent course recommendation system that delivers customized educational routes to students. The system solves challenges of effective course recommendations through employing various recommendation approaches, such that the drawbacks of one method are covered up for by another. The recommendation system implements three recommendation strategies through content-based filtering for description and material analysis as well as collaborative filtering for user collective preferences followed by knowledge-based filtering for structured course prerequisites and learning objectives. We build transparency in the system along with user trust using Large Language Model (LLM)-based explanations which provide justification details for the recommended items. The system responds to the essential requirement of explainable intelligent systems. Our system receives evaluation through an extensive dataset that includes user behaviour records with additional information about courses and professional objectives. The system proves its ability to create appropriate recommendations while ensuring diversity and personalization thus solving the typical challenges of course discovery for new subjects and recommendation uniformity. The F1-score of the model comes out to be around 0.8, which is higher than the average ratings for traditional systems, which usually lie in the range of 0.57–0.72. These advances represent critical milestones in developing smart educational platforms which provide individualized support for lifelong learning journeys.