Home // International Journal On Advances in Software, volume 17, numbers 3 and 4, 2024 // View article

Beyond Stars: Enriching Restaurant Reviews with Interactive Follow-Up Analysis

Authors:
Kaho Mizobata
Ryosuke Yamanishi

Keywords: Follow-up interaction; computational approach for food and eating activities; Large Language Model-supported system.

Abstract:
This paper proposes a follow-up interaction system designed to enhance restaurant reviews and evaluates its effectiveness through empirical analysis. Restaurant reviews serve as a critical source of information for customers when selecting dining options and significantly influence a restaurant’s reputation and patronage. However, many reviews are missing some points to be reviewed, often omitting important aspects of the dining experience. To address this issue, this study introduces a system leveraging ChatGPT to identify missing elements in reviews and prompt reviewers to include them through follow-up interactions, thereby enriching the content of reviews. The experiment observed participants as they refined their reviews using the system’s feedback. We analyzed the originally described elements, the system-identified absent elements, and the elements added after follow-up interactions. The results demonstrated that follow-up interactions effectively increased the amount of information in reviews and ensured a comprehensive coverage of multiple perspectives, including food, restaurant environment, and reviewer experiences. Additionally, we conducted statistical analyses to examine co-occurrence patterns between review elements and assess the fairness of the system’s suggestions for absent elements. The findings highlighted the potential of this system to improve the quality of user-generated content. We believe that it would enable consumers to access detailed and reliable reviews while providing restaurants with actionable customer feedback to enhance their services.

Pages: 239 to 248

Copyright: Copyright (c) to authors, 2024. Used with permission.

Publication date: December 30, 2024

Published in: journal

ISSN: 1942-2628