基于LDA的可穿戴设备在线评论主题挖掘研究

发布时间：2018-02-21 14:57

  本文关键词： 智能手环 可穿戴设备 LDA 用户满意度 在线评论 主题挖掘　出处：《信息资源管理学报》2017年03期 　论文类型：期刊论文

【摘要】：本文利用文本挖掘方法分析了智能手环用户评论中的热点主题,以了解用户对不同品牌智能手环关注点差异以及用户关注主题随时间的变化。首先,利用LDA模型对文本进行主题分类,了解不同品牌用户关注主题的排序;然后对不同时段各品牌的评论信息进行分析,了解不同用户群体对产品功能关注度的变化;最后分析用户对不同品牌手环整体满意度的差异,并结合评论文本信息分析用户使用时长与满意度之间的关系。本研究发现对于所有品牌,用户均关注产品功能和外观,以及在使用中发现的产品问题和感受,但是不同品牌关注主题排序各有差异。就用户满意度而言,随着使用时间的增长,用户对小米手环的满意度日渐降低,而Fitbit和华为手环的用户满意度与使用时间呈正U形曲线。
[Abstract]:In this paper, a text mining method is used to analyze the hot topics in user reviews of smart bracelet, in order to understand the difference of user's attention to different brands of smart bracelet and the change of user's concern over time. The LDA model is used to classify the theme of the text to understand the ranking of the topics concerned by the users of different brands, and then to analyze the comments of different brands in different periods to understand the changes of the attention of different user groups to the function of the product. Finally, the paper analyzes the difference of users' overall satisfaction with different brands of bracelet, and analyzes the relationship between the length of time used by users and their satisfaction with the commentary text information. This study finds that for all brands, the users are concerned about the function and appearance of the products. And the problems and feelings found in the use of the products, but different brands pay attention to the theme of different ranking. In terms of user satisfaction, as the use of time, the user's satisfaction with the Xiaomi bracelet gradually decreased, Fitbit and Huawei bracelet user satisfaction and use time positive U-shaped curve.
【作者单位】： 武汉大学信息管理学院;武汉大学电子商务研究与发展中心;
【基金】：国家自然科学基金项目“内容关系互动下的在线医疗社区用户行为演化研究”(71573197)和“物联网环境下组织体系架构建模、行为分析与优化设计“(71531009)的研究成果之一
【分类号】：G252;TP368.33
，

本文编号：1522172