How review sentiment influences ratings in Generative AI applications: Insights from VADER and LDA analysis

IF 13.1 1区管理学 Q1 BUSINESS

Journal of Retailing and Consumer Services Pub Date : 2025-10-04 DOI:10.1016/j.jretconser.2025.104560

Yuebin Meng , Sangchul Park , Geumchul Um

{"title":"How review sentiment influences ratings in Generative AI applications: Insights from VADER and LDA analysis","authors":"Yuebin Meng , Sangchul Park , Geumchul Um","doi":"10.1016/j.jretconser.2025.104560","DOIUrl":null,"url":null,"abstract":"<div><div>Generative AI applications have emerged as a crucial medium for users' daily intelligent interactions with user sentiment tendencies serving as a vital factor influencing product optimization and market competition. The purpose of this study was to explore the relationship between the title and review sentiment and rating and its boundary conditions. To this end, this study collected 100,010 user reviews from nine major generative AI applications in the U.S. App Store, using integrated VADER sentiment analysis scores with LDA topic weights. The results showed that the sentiment scores of reviews and titles have a significant positive effect on user ratings, respectively. Moreover, this study verified (1) the negative moderating effect of the review length and title length and that (2) the value-related topics enhance the review sentiment score, while technology-related and functionality-related topics weaken the review sentiment score. These findings provide empirical evidence for product managers and operations teams of generative AI platforms for optimizing their products and securing a competitive edge in the market.</div></div>","PeriodicalId":48399,"journal":{"name":"Journal of Retailing and Consumer Services","volume":"88 ","pages":"Article 104560"},"PeriodicalIF":13.1000,"publicationDate":"2025-10-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Retailing and Consumer Services","FirstCategoryId":"91","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S096969892500339X","RegionNum":1,"RegionCategory":"管理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"BUSINESS","Score":null,"Total":0}

引用次数: 0

Abstract

Generative AI applications have emerged as a crucial medium for users' daily intelligent interactions with user sentiment tendencies serving as a vital factor influencing product optimization and market competition. The purpose of this study was to explore the relationship between the title and review sentiment and rating and its boundary conditions. To this end, this study collected 100,010 user reviews from nine major generative AI applications in the U.S. App Store, using integrated VADER sentiment analysis scores with LDA topic weights. The results showed that the sentiment scores of reviews and titles have a significant positive effect on user ratings, respectively. Moreover, this study verified (1) the negative moderating effect of the review length and title length and that (2) the value-related topics enhance the review sentiment score, while technology-related and functionality-related topics weaken the review sentiment score. These findings provide empirical evidence for product managers and operations teams of generative AI platforms for optimizing their products and securing a competitive edge in the market.

查看原文本刊更多论文

评论情绪如何影响生成式AI应用中的评分：来自VADER和LDA分析的见解

生成式人工智能应用已经成为用户与用户情绪趋势进行日常智能交互的重要媒介，成为影响产品优化和市场竞争的重要因素。本研究的目的是探讨标题与评论情绪和评分之间的关系及其边界条件。为此，本研究从美国App Store的9个主要生成式AI应用中收集了10010条用户评论，使用集成的VADER情感分析得分和LDA主题权重。结果表明，评论和标题的情感得分分别对用户评分有显著的正向影响。此外，本研究验证了(1)评论长度和标题长度的负向调节作用；(2)价值相关主题提高了评论情绪得分，而技术和功能相关主题削弱了评论情绪得分。这些发现为生成式人工智能平台的产品经理和运营团队优化产品和确保市场竞争优势提供了经验证据。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Journal of Retailing and Consumer Services BUSINESS-

CiteScore

20.40

自引率

14.40%

发文量

340

审稿时长

20 days

期刊介绍： The Journal of Retailing and Consumer Services is a prominent publication that serves as a platform for international and interdisciplinary research and discussions in the constantly evolving fields of retailing and services studies. With a specific emphasis on consumer behavior and policy and managerial decisions, the journal aims to foster contributions from academics encompassing diverse disciplines. The primary areas covered by the journal are: Retailing and the sale of goods The provision of consumer services, including transportation, tourism, and leisure.