Kaho Mizobata, Ryosuke Yamanishi: Analysis and Enrichment of Description in Restaurant Review through Follow-Up Interaction, The International Conference on Advances in Computer-Human Interactions 2024, (to appear), 2024年5月

This paper proposes a framework to enrich restaurant reviews by providing follow-up questions to reviewers about absent elements in their original reviews. Utilizing ChatGPT, we investigated enhancing the detail and organization of reviews by examining 26 participants’ interactions across food, environment, and user experience. The results suggested that the follow-up interaction encouraged more informative reviews by highlighting omitted details. Especially, it effectively increases mentions of the restaurant’s atmosphere and personal experiences alongside food descriptions. This approach offers insights into factors influencing review content, such as review writing experience and dining context. We believe that the findings will be helpful for customers as a guide to writing reviews and suggest the effectiveness of follow-up interaction in writing reviews.