A decision support system for comments-adjusted ranking of hotels
提出一种基于云模型的酒店排名方法，通过调整用户评分和量化评论文本，生成推荐列表，并用TripAdvisor数据验证，帮助游客更准确选择酒店。
Increasing online reviews provide a powerful support tool to obtain a relevant hotel selection for tourists. Some reviews seem to have similar contents but their latent meanings are discrepant, due to user preferences. It has a negative effect on the recommendation result without a well processing. To solve this issue, this study develops a cloud-based ranking method to provide hotel recommendation lists for tourists. The method uses the distribution of historical user ratings to adjust original user ratings for minimizing the understanding bias among reviews, and utilizes sentiment analysis and cloud models to quantify text comments. After these, both of them are merged into a holistic score to generate a hotel recommendation list. Our method is validated through a case study of 87,735 users and 835,462 reviews from TripAdvisor.com. Further, our research findings suggest that potential hotel stayers should realize to lower their expectations of a good stay from online reviews, and trust more online reviews of couple and friend travelers.