Support for Restaurant Comparison by Presenting Feature Words in Review Sentences on Gourmet Websites
Anna Sakai (AY 2021)
Purpose: Gourmet websites are one of the information sources used to obtain information about restaurants. By using a gourmet restaurant website, users can mainly obtain basic information about restaurants, such as opening hours, location, and menu items, as well as user reviews of the restaurants. Therefore, gourmet sites have the advantage of not only providing basic information about restaurants, but also allowing users to compare multiple restaurants by referring to the reviews of users who have actually used them. However, the number of restaurants listed on a gourmet restaurant website is huge, and the number of words in each review is also large, so users have to spend a lot of time to compare multiple restaurants. Therefore, the purpose of this study was to facilitate the comparison of restaurants in a short time by extracting characteristic words of restaurants from reviews on a gourmet restaurant website and presenting them to users.
Methodology: A feature word presentation screen was created based on the search results screen of Eat-Log, one of the gourmet sites. Feature words were manually selected based on words related to ingredients and flavors that frequently appear in word-of-mouth documents, as well as frequently appearing dish names and prices. To verify the usefulness of the created system, an evaluation experiment was conducted with 20 participants. First, a survey was administered about the participants' experience in using gourmet food websites. Next, the participants were asked to select restaurants in a scenario where they would search for restaurants in an unfamiliar area, using both the screen of the conventional system "Eating Log" and the proposed system. After selecting the restaurants, a questionnaire was administered regarding the satisfaction with the two systems in selecting the restaurants and the satisfaction with the systems used. In addition, the work screens during the restaurant selection and the time taken for the selection were recorded.
RESULTS: The results of the experiment showed no significant differences in the time taken to compare restaurants, the number of screen transitions, and the level of satisfaction. However, in terms of satisfaction, positive opinions were obtained from the free responses regarding the proposed method. In particular, positive responses were found for the items of smoothness of selection and ease of use of the system. In the questionnaire survey conducted at the end of the experiment, 70% of the participants answered that the proposed method was easier to use, and positive opinions were also obtained regarding the presentation of feature words.
Conclusion: Although there was no significant difference between the conventional and proposed methods in terms of selection time or satisfaction, it was confirmed that the presentation of feature words can contribute to improving the ease of comparing and selecting restaurants. In addition, a feature word search function was desired in the experiment. In addition, it would be possible to make the presentation of feature words more effective by implementing a function that allows the presentation of both feature words and part of the word-of-mouth, or by changing the presentation method according to the user's needs.
(Translated by DeepL)