A Hotel Recommendation System Based on Reviews: What Do You Attach Importance To?

Abstract

In looking for a hotel, it is common to access a list of hotels matching a query which is arranged in a descending order of the average evaluation value. Since such a list does not reflect the preference of users, for many inexperienced users, it takes too long to determine a hotel. In this paper, we focus on the evaluation values given by contributors whose preferences are similar to the user's preference. Such evaluation values may be more highly credible for the user. We proposes a method to extract the preference of review contributors from a collection of reviews. The extracted preferences are used for the hotel recommendation in such a way that the evaluation value given by a contributor to have preference similar to the user is given larger weight. The result of questionnaire-based evaluations indicates that our proposed method can recommend hotels that matches the user preference.