Scholarlink

<p>User engagement has been improved by using recommender systems, which are essential for giving user recommendations. Matrix factorization (MF), one of the traditional approaches, has shown a promising work in capturing user preferences, especially in certain areas like movie recommendation with the help of MovieLens, which are datasets. Nevertheless, there is always a need to optimize long-term user engagement to overcome the challenges like large action space in recommender systems and the recent advancements like deep reinforcement learning (DRL) have shown a much greater role in addressing the same. This study aims to explore the effectiveness of hybrid approaches, wherein our innovative HybridFlicks recommender system is proposed, which combines the strengths of MF-based collaborative filtering with content-based filtering using TF-IDF and Cosine similarity, to improve the accuracy of the recommendations provided to the user, by focusing on enhancing the recall value and F1 score.</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

HybridFlicks: integrating collaborative filtering and content-based methods for enhanced movie recommendations

  • 期刊论文

  • 发表日期：

    2025年5月14日

查看全文
  • Prabakar Dakshinamoorthy,
  • Arvindh Murugan,
  • Sohan Vijayabhaskar,
  • Bharani Kumar Thirumal

摘要

User engagement has been improved by using recommender systems, which are essential for giving user recommendations. Matrix factorization (MF), one of the traditional approaches, has shown a promising work in capturing user preferences, especially in certain areas like movie recommendation with the help of MovieLens, which are datasets. Nevertheless, there is always a need to optimize long-term user engagement to overcome the challenges like large action space in recommender systems and the recent advancements like deep reinforcement learning (DRL) have shown a much greater role in addressing the same. This study aims to explore the effectiveness of hybrid approaches, wherein our innovative HybridFlicks recommender system is proposed, which combines the strengths of MF-based collaborative filtering with content-based filtering using TF-IDF and Cosine similarity, to improve the accuracy of the recommendations provided to the user, by focusing on enhancing the recall value and F1 score.