Scholarlink

<p>Recommendation systems play a crucial role in the tourism industry, helping users navigate the overwhelming amount of available information. The existing hotel recommendation platforms often overlook personalized user preferences, leading to generic suggestions. This research presents an advanced Feature and Content Aware Service Recommender (FCASR) designed to enhance personalized hotel recommendations. FCASR constructs detailed user profiles by integrating content analysis and sentiment evaluation of user feedback, capturing individual preferences for more precise recommendations. Built on a Hadoop-based distributed framework, it utilizes big data processing to ensure scalability and efficiency. Experiments on a real-world TripAdvisor dataset demonstrate that FCASR significantly improves recommendation accuracy through opinion mining and optimized similarity measures. Its efficient MapReduce implementation reduces computational overhead, making it more time-efficient than traditional keyword-aware approaches. The results confirm that FCASR outperforms existing techniques, offering a robust, scalable, and high-quality service recommendation solution. FCASR ESC achieves 9–30% lower MAE than baselines, while remaining competitive with DistilBERT at a fraction of the computational cost.</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

FCASR: a feature and content-aware service recommender for personalized hotel recommendations using big data

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月21日

查看全文
  • Saad Azhar Saeed,
  • Zareen Alamgir

摘要

Recommendation systems play a crucial role in the tourism industry, helping users navigate the overwhelming amount of available information. The existing hotel recommendation platforms often overlook personalized user preferences, leading to generic suggestions. This research presents an advanced Feature and Content Aware Service Recommender (FCASR) designed to enhance personalized hotel recommendations. FCASR constructs detailed user profiles by integrating content analysis and sentiment evaluation of user feedback, capturing individual preferences for more precise recommendations. Built on a Hadoop-based distributed framework, it utilizes big data processing to ensure scalability and efficiency. Experiments on a real-world TripAdvisor dataset demonstrate that FCASR significantly improves recommendation accuracy through opinion mining and optimized similarity measures. Its efficient MapReduce implementation reduces computational overhead, making it more time-efficient than traditional keyword-aware approaches. The results confirm that FCASR outperforms existing techniques, offering a robust, scalable, and high-quality service recommendation solution. FCASR ESC achieves 9–30% lower MAE than baselines, while remaining competitive with DistilBERT at a fraction of the computational cost.