基于在线评论和量化 Kano 模型的用户需求识别研究

摘要：

【目的】通过对用户在线评论数据进行分析，帮助企业准确识别并深入理解用户需求。【方法】构建了基于在线评论和量化Kano模型的用户需求识别方法，首先运用LDA模型对用户在线评论进行主题聚类，然后分别利用BosonNLP和文本挖掘技术计算用户满意度和用户关注度指标，最后基于这2个指标构建量化Kano模型实现对用户不同需求的识别与分类。【结果】以小米社区为案例进行实验。结果表明，用户在不同产品属性的关注度和满意度上存在显著差异，屏幕大小和应用生态兼容性属于必备型需求，电池续航属于魅力型需求，系统流畅属于期望型需求，拍照清晰度和外观颜色属于无差异型需求。【结论】本方法突破了传统调研主观性较强的局限，为企业提升用户满意度和优化资源配置提供了数据驱动的决策依据。

Abstract:

[Objective] Through the analysis of user online review data, the study helps enterprises to accurately identify and understand user requirements. [Methods] A user requirement identification method based on online reviews and quantitative Kano model is constructed. Firstly, the LDA model is used to cluster the topics of users’ online reviews. Then, BosonNLP and text mining techniques are used to calculate user satisfaction and attention indicators, respectively. Finally, the quantitative Kano model is constructed to identify and classify different requirements of users. [Results]. This paper takes the Xiaomi community as a case study. The experimental results show that there are significant differences in the attention and satisfaction of users in different product attributes, such as screen size and application ecological compatibility are must-be requirement, battery is an attractive requirement, system smoothness is a one-dimensional requirement, and image clarity and color are indifferent requirements. [Conclusion] This method breaks through the limitations of traditional research with strong subjectivity, and provides a data-driven decision-making basis for enterprises to improve user satisfaction and optimize resource allocation. 

引用格式：曲壮，方兴华，宋明顺.基于在线评论和量化Kano模型的用户需求识别研究[J].标准科学，2025(6)：41-50.

基金项目：本文受教育部人文社会科学研究青年基金项目“生产商主导型C2M模式下产品优化配置研究：基于异质性顾客需求视角”（项目编号：22YJC630022）；浙江省基本科研业务费资助项目“考虑顾客需求异质性的质量设计方法研究”（项目编号：2022YW29）；浙江省软科学重点项目“浙江省数字经济创新提质发展思路研究”（项目编号：2024C25046）资助。 曲壮，硕士研究生，研究方向为质量管理、用户需求分析。 

作者简介：方兴华，博士，副教授，研究方向为质量管理、产品开发理论。 宋明顺，博士，教授，研究方向为标准化与质量管理。