基于LDA的电商平台用户评论挖掘与情感分析研究——以京东商城App为例  

Research on user review mining and sentiment analysis of E-commerce platform based on LDA:taking Jingdong Mall App as an example

作　　者：杜利明 郭文艳[2] 崔蕾 王凤英 DU Liming;GUO Wenyan;CUI Lei;WANG Fengying(School of Information Engineering,Suqian University,Suqian 223800,China;School of Computer Science and Engineering,Shenyang Jianzhu University,Shenyang 110000,China)

机构地区：[1]宿迁学院信息工程学院,江苏宿迁223800 [2]沈阳建筑大学计算机科学与工程学院,辽宁沈阳110000

出　　处：《江苏科技信息》2024年第12期125-129,共5页Jiangsu Science and Technology Information

基　　金：宿迁学院京东学院开放基金项目,项目名称:基于京东用户评论挖掘的市场情报分析模型研究,项目编号:2022JDXM13。

摘　　要：用户评论文本挖掘与分析在多个领域具有重要实际应用价值。文章选取京东商城用户评论数据集作为研究对象,运用多种方法对其进行深入的数据挖掘与分析。首先,通过TF-IDF提取关键词揭示评论的核心主题,进而通过分析高频词了解用户对京东商城服务的关注点和整体评价。其次,采用情感分析技术对评论文本进行情感倾向性分类,旨在判断评论的情绪色彩,为京东商城的产品改进和市场定位提供有益参考。最后,借助LDA主题模型对评论文本进行主题剖析,挖掘出评论中的隐性主题和话题分布,进一步揭示用户对产品或服务的不同观点和需求,从而为京东商城提供针对性的改进策略和意见。User comment text mining and analysis have important practical application value in multiple fields.The article selects the user comment dataset of Jingdong Mall as the research object,and uses various methods to conduct in-depth data mining and analysis on it.Firstly,extract keywords through TF-IDF to reveal the core theme of the comment,and then analyze high-frequency words to understand the user’s focus and overall evaluation of Jingdong Mall’s services.Secondly,sentiment analysis technology is used to classify the emotional tendencies of comment texts,aiming to determine the emotional color of comments and provide useful references for product improvement and market positioning of Jingdong Mall.Finally,using the LDA topic model to analyze the theme of the comment text,the implicit themes and topic distribution in the comment are excavated,further revealing the different views and needs of users on the product or service,and providing targeted improvement strategies and opinions for Jingdong Mall.

关 键 词：LDA 用户评论 文本挖掘 情感分析 

分 类 号：F724.6[经济管理—产业经济] F426.8