基于BERTopic和提示学习的AIGC类移动应用用户需求研究  

Research on User Requirement of AIGC Mobile Applications Based on BERTopic and Prompt Learning

作　　者：强甜 Qiang Tian(School of Public Administration,Sichuan University,Chengdu Sichuan 610065)

机构地区：[1]四川大学公共管理学院,四川成都610065

出　　处：《情报探索》2025年第1期79-86,共8页Information Research

摘　　要：[目的/意义]人工智能生成内容的发展促进了用户服务转型。移动应用是AIGC技术的重要载体,评论主题识别和用户需求分析能为本土AIGC类APP的推广提供支持。同时,大语言模型提示学习也能提升评论处理的效率。[方法/过程]豆包APP是国内典型的AIGC移动应用。本研究从七麦数据中获取了8434条评论数据,采用GLM-4大语言模型提示学习识别有效语料,运用BERTopic模型提取需求主题。参考国际软件质量标准,深入理解用户软件需求。[结果/结论]从评论中识别出13个与用户软件需求相关的核心主题,并提出用户功能性需求、性能效率需求、可使用性需求和可靠性需求。最后从功能拓展、人机交互、运营策略和软件稳定性四个方面提供了发展建议。[Purpose/significance]The development of artificial intelligence generated has promoted the transformation of user services.Mobile applications is an important carrier of AIGC technology.Comment topic identification and user demand analysis can support for the promotion of local AIGC APP.At the same time,large language model prompt learning can also improve the efficiency of comment processing.[Method/process]Beab bag APP is a typical AIGC mobile application in China.This study collects 8434 comments from Qimai data,employing the GLM-4 language model to prompt learning to identify effective corpus,and applying BERTopic model to extract demand topics.This artical refers to the international software quality standards,in-depth understanding of user software requirements.[Result/conclusion]From the comments,it identifies 13 core topics related to user software requirements,user functional requirements,performance efficiency requirements,usability requirements and reliability requirements.Finally,it provides development suggestions from four aspects:function expansion,human-computer interaction,operation strategy and software stability.

关 键 词：AIGC 移动应用 用户需求 BERTopic 提示学习 

分 类 号：G252[文化科学—图书馆学] G353.1