摘要
[目的/意义]为克服用户功能需求挖掘过程中仅以消极评论中高频关键词作为分析依据,导致用户功能需求定位不准确的问题,提出基于结构主题模型的用户功能需求研究框架。[方法/过程]首先,采用结构主题模型(STM)方法,通过对比积极和消极评论提取具有消极倾向的功能主题,并对其中的词语和高相关性评论进行可视化,挖掘用户功能需求;将其与传统词频统计方法进行对比,分析STM方法的适用性及优势;然后,探索分析各功能需求间的关联关系以及产品价格变化对不同功能需求的影响程度,为提高产品的整体性能和科学产品定位提供支撑;最后,以老年可穿戴式跌倒报警产品为例验证该研究框架的有效性。[结果/结论]结果表明:STM克服了传统词频统计方法的片面性缺陷,该框架将积极评论与消极评论同时纳入主题提取过程,通过提取在消极评论中远高于积极评论中的高频关键词挖掘用户不满意的真正原因,提高了用户功能需求定位的准确性,为企业及相关研发部门制定合理研发战略提供有利指导。
[Purpose/Significance]That the user’s functional requirement mining process only uses the high-frequency keywords in the negative comments as the analysis basis leads to the inaccurate positioning of the user’s functional requirements.In order to overcome the problem,this paper proposes a user functional requirements research framework based on the Structural Topic Model(STM).[Method/Process]Firstly,the STM method is used to extract the functional topics with negative tendencies by comparing positive and negative comments and then visualizing the words and highly relevant comments in them to mine the functional needs of users.In order to analyze its applicability and advantage,the STM method is compared with the traditional word frequency statistics method.Secondly,the relationship among the functional requirements and the degree of impact of changes in product prices on different functional requirements are explored and analyzed to provide support for improving the overall performance of the product and scientific product positioning.Finally,an elderly wearable fall alarm product is used as an example to verify the validity of the research framework.[Result/Conclusion]The results show that STM overcomes the one-sidedness defect of the traditional word frequency statistical method.The framework incorporates both positive and negative reviews into the topic extraction process.It extracts high-frequency keywords that are much higher in negative reviews than in positive reviews to discover the real cause of user dissatisfaction,which not only improves the accuracy of user functional requirements positioning,but also provides favorable guidance for enterprises and related R&D departments to formulate reasonable R&D strategies.
作者
苗红
王艳
黄鲁成
吴菲菲
李欣
Miao Hong;Wang Yan;Huang Lucheng;Wu Feifei;Li Xin(School of Economics and Management,Beijing University of Technology,Beijing 100124)
出处
《情报杂志》
CSSCI
北大核心
2020年第7期134-141,共8页
Journal of Intelligence
基金
国家社会科学基金重大项目“依靠科技创新应对人口老龄化跨学科研究”(编号:17ZDA119)研究成果之一。
关键词
用户功能需求
结构主题模型
在线评论
情感倾向
老年可穿戴式跌倒报警产品
user functional requirements
Structural Theme Model
online reviews
sentiment analysis
elderly wearable fall alarm product
