图书情报工作

图书情报工作 >

2025 , Vol. 69 >Issue 17: 130 - 142

DOI: https://doi.org/10.13266/j.issn.0252-3116.2025.17.011

研究论文

产品评论中的“情绪—原因—特征”用户需求网络构建与分析

  • 靳健 ,
  • 吴倩 ,
  • 秦淑英
展开
  • 1 北京师范大学政府管理学院 北京 100875;
    2 山西大学图书馆 太原 030006
靳健，教授，博士，硕士生导师；吴倩，硕士研究生；秦淑英，助理馆员，硕士研究生，通信作者，E-mail：qsy2372167673@163.com。

收稿日期: 2024-06-11

  修回日期: 2024-09-18

  网络出版日期: 2025-08-20

基金资助

本文系国家自然科学基金面上项目“利企便民场景中政策信息语义化与智能推送的方法研究”（项目编号：72374030）研究成果之一。

收起

Construction and Analysis of “Emotion-Cause-Feature” User Requirement Network in Product Reviews

  • Jin Jian ,
  • Wu Qian ,
  • Qin Shuying
Expand
  • 1 School of Government, Beijing Normal University, Beijing 100875;
    2 Shanxi University Library, Taiyuan 030006
Jin Jian, Professor, PhD, Master’s Supervisor; Wu Qian, Master’s Candidate; Qin Shuying, Assistant Librarian, Master’s Candidate, Corresponding Author, E-mail: qsy2372167673@163.com.

Received date: 2024-06-11

  Revised date: 2024-09-18

  Online published: 2025-08-20

Supported by

This work is supported by the National Natural Science Foundation of China, titled “An Investigation on Policy Semantic Annotation and Intelligent Recommendation in the Scenario of Government-Enterprise Encouragement and Government-Citizen Interaction” (Grant No. 72374030).

Fold

摘要

[目的/意义] 通过识别和分析用户评论中的情绪原因，深入挖掘用户需求，以支持用户需求导向的产品设计和优化。[方法/过程] 将用户情感细化为用户情绪，首先根据预训练模型—微调范式，对预训练语言模型T5进行微调，再用训练好的模型抽取产品评论中与用户情绪对应的原因子句；紧接着从原因子句中抽取产品描述微原因和使用场景微原因，并构成两类用户情绪原因摘要；最后提出“情绪—原因—特征”用户需求网络模型，将先前步骤抽取的各因素结合起来进行可视化，在此基础上展开用户需求分析。同时以公开的手机产品评论数据集为例，验证模型的有效性。[结果/结论] 结果表明，相较于传统情绪识别任务，细粒度更高的情绪原因识别能够更简明、有效地捕捉用户需求，反映消费者对产品的喜好、不满和期望，为产品设计者提供优化思路，帮助市场营销人员把握产品宣传方向。

关键词： 用户需求; 情绪原因识别; 评论摘要; 需求网络; 产品评论

本文引用格式

靳健 , 吴倩 , 秦淑英 . 产品评论中的“情绪—原因—特征”用户需求网络构建与分析[J]. 图书情报工作, 2025 , 69(17) : 130 -142 . DOI: 10.13266/j.issn.0252-3116.2025.17.011

Abstract

[Purpose/Significance] Identifying and analyzing emotional causes in user reviews can deeply uncover user needs, and support user-centered product design and optimization. [Method/Process] In this study, user emotions were fine-grained, focusing on words that accurately express users’ emotions. Firstly, the pre-trained language model T5 was fine-tuned, and was used to extract the cause clauses corresponding to user emotions in product reviews. Secondly, the product description micro-cause and usage scenario micro-cause were extracted to generate two types of comment summaries correspondingly. Finally, then this paper proposed an “emotion-cause-feature” user requirement network model, which combined the factors extracted in the previous steps for visualization, and then carried out user demand analysis on this basis. This study took the public mobile phone product review data set as an example to verify the effectiveness of the model. [Result/Conclusion] The results show that compared with traditional emotion recognition tasks, finer-grained emotion cause identification can capture user needs more concisely and effectively. It can also better reflect consumer’s preferences, dissatisfaction, and expectations of products, providing optimization ideas for product designers and aiding marketers in determining the direction of product promotion.

Key words： user needs; emotion cause identification; comment summarization; requirement network; product reviews

参考文献

[1] 李汝鹏, 魏巍, 周峰, 等. 面向客户需求的产品优化设计方法研究[J]. 中国工程科学, 2018, 20(2): 33-41. (LI R P, WEI W, ZHOU F, et al. Research on product optimization design method to respond rapidly to customer requirements[J]. Strategic study of CAE, 2018, 20(2): 33-41.)
[2] 贾丹萍, 靳健, 耿骞, 等. 感性工学视角下的用户需求挖掘研究[J]. 情报学报, 2020, 39(3): 308-316. (JIA D P, JIN J, GENG Q, et al. A Kansai engineering integrated approach for customer-needs mining from online product reviews[J]. Journal of the China Society for Scientific and Technical Information, 2020, 39(3): 308-316.)
[3] 王克勤, 毋凤君. 面向产品设计改进的在线评论挖掘[J]. 计算机工程与应用, 2019, 55(19): 235-245, 252. (WANG K Q, WU F J. Online reviews mining for product design improvement[J]. Computer engineering and applications, 2019, 55(19): 235-245, 252.)
[4] JIN J, LIU Y, Ji P, et al. Understanding big consumer opinion data for market-driven product design[J]. International journal of production research, 2016, 54(10): 3019-3041.
[5] 李江泳, 张伟. 在线评论数据挖掘视角下的产品改良设计研究[J]. 包装工程, 2021, 42(6): 135-141. (LI J Y, ZHANG W. Product improvement design from the perspective of online comment data mining[J]. Packaging engineering, 2021, 42(6): 135-141.)
[6] TIMOSHENKO A, HAUSER J R. Identifying customer needs from user-generated content[J]. Marketing science, 2019, 38(1): 1-20.
[7] 沈超, 王安宁, 方钊, 等. 基于在线评论数据的产品需求趋势挖掘[J]. 中国管理科学, 2021, 29(5): 211-220. (SHEN C, WANG A N, FANG Z, et al. Trend mining of product requirements from online reviews[J]. Chinese journal of management science, 2021, 29(5): 211-220.)
[8] 蔡庆平, 马海群. 基于Word2Vec和CNN的产品评论细粒度情感分析模型[J]. 图书情报工作, 2020, 64(6): 49-58. (CAI Q P, MA H Q. A fine-grained sentiment analysis model for product reviews based on Word2Vec and CNN[J]. Library and infor-mation service, 2020, 64(6): 49-58.)
[9] 韦景竹, 操慧子, 张乐乐. 基于在线评论的公共文化云活动用户需求研究[J]. 图书情报工作, 2022, 66(9): 66-73. (WEI J Z, CAO H Z, ZHANG L L. Research on users’ needs of public culture cloud activities based on online comments[J]. Library and information service, 2022, 66(9): 66-73.)
[10] 李贺, 曹阳, 沈旺, 等. 基于LDA主题识别与Kano模型分析的用户需求研究[J]. 情报科学, 2021, 39(8): 3-11, 36. (LI H, CAO Y, SHEN W, et al. User demand based on LDA subject identification and Kano model analysis[J]. Information science, 2021, 39(8): 3-11, 36.)
[11] 涂海丽, 唐晓波, 谢力. 基于在线评论的用户需求挖掘模型研究[J]. 情报学报, 2015, 34(10): 1088-1097. (TU H L, TANG X B, XIE L. Research on user needs mining model based on online reviews[J]. Journal of the China Society for Scientific and Technical Information, 2015, 34(10): 1088-1097.)
[12] 马超, 李纲, 陈思菁, 等. 基于多模态数据语义融合的旅游在线评论有用性识别研究[J]. 情报学报, 2020, 39(2): 199-207. (MA C, LI G, CHEN S J, et al. Research on usefulness recognition of tourism online reviews based on multimodal data seman-tic fusion[J]. Journal of the China Society for Scientific and Technical Information, 2020, 39(2): 199-207.)
[13] CHEN Y, LEE Y M, LI S S, et al. Emotion cause detection with linguistic constructions[C]//Proceedings of the 23rd Int Conf on Computational Linguistics. Beijing: Tsinghua University Press, 2010: 179-187.
[14] GUI L, WU D Y, XU R F, et al. Event-driven emotion cause extraction with corpus construction[C]//Proceedings of the 2016 Conf on Empirical Methods in Natural Language Processing. Stroudsburg: ACL, 2016: 1639-1649.
[15] LEE Y M, CHEN Y, HUANG C R. A text-driven rule-based system for emotion cause detection[C]//Proceedings of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT). Stroudsburg: ACL, 2010: 45-53.
[16] 李逸薇, 李寿山, 黄居仁, 等. 基于序列标注模型的情绪原因识别方法[J]. 中文信息学报, 2013, 27(5): 93-99. (LEE Y M, LI S S, HUANG J R, et al. Detecting emotion cause with sequence labeling model[J]. Journal of Chinese information pro-cessing, 2013, 27(5): 93-99.)
[17] CHENG X Y, HOU W J, CHEN Y. Hierarchical convolution neural network for emotion cause detection on mi-croblogs[C]//The 27th International Conference on Artificial Neural Networks (ICANN). Switzerland: Springer, 2018: 115-122.
[18] 夏林旭, 刘茂福, 胡慧君. 基于AM-BiLSTM模型的情绪原因识别[J]. 武汉大学学报(理学版), 2019, 65(3): 276-282. (XIA L X, LIU M F, HU H J. Emotion cause recognition based on AM-BiLSTM model[J]. Journal of Wuhan University (natural sci-ence edition), 2019, 65(3): 276-282.)
[19] 万漪萍, 赵岩, 赵留彦. 投资者情绪与证券投资收益：机器学习对网络社交媒体的文本分析[J]. 金融市场研究, 2024(5): 65-77. (WAN Y P, ZHAO Y, ZHAO L Y. Stock returns and investor sentiment: machine learning for textual analysis of online social media[J]. Financial market research, 2024(5): 65-77.)
[20] 耿晓媛. 投资者情绪的识别与统计——基于非结构数据的分析[J]. 科学决策, 2023(11): 156-169. (GENG X Y. Identifica-tion and statistics of investor emotions: analysis based on unstructured data[J]. Scientific decision making, 2023(11): 156-169.)
[21] 刘柏嵩, 赵福青. 基于微观点的产品评论微摘要研究[J]. 情报学报, 2015, 34(9): 970-977. (LIU B S, ZHAO F Q. Mi-crosummary of product reviews based on micropinion[J]. Journal of the China Society for Scientific and Technical Information, 2015, 34(9): 970-977.)
[22] LOVINGER J, VALOVA I, CLOUGH C. Gist: general integrated summarization of text and reviews[J]. Soft Computing, 2019, 23(5SI): 1589-1601.
[23] 林莉媛, 王中卿, 李寿山, 等. 基于评论质量的多文档文本情感摘要[J]. 中文信息学报, 2015, 29(4): 33-39. (LIN L Y, WANG Z Q, LI S S, et al. Quality-based multi-document opinion summarization[J]. Journal of Chinese information processing, 2015, 29(4): 33-39.)
[24] YANG M, QU Q, SHEN Y, et al. Aspect and sentiment aware abstractive review summarization[C]//The 27th International Conference on Computational Linguistics. Stroudsburg: ACL, 2018: 1110-1120.
[25] 李红莲, 陈浩天, 张乐, 等. 融合情感-主题双通道信息的评论摘要生成模型[J]. 数据分析与知识发现, 2024, 8(6): 30-43. (LI H L, CHEN H T, ZHANG L, et al. A review summary generation model with emotion-topic dual-channel information[J]. Data analysis and knowledge discovery, 2024, 8(6): 30-43.)
[26] 陶嘉鸿, 卢永美, 何东, 等. 融合用户属性交互的个性化评论摘要生成[J]. 小型微型计算机系, 2023, 44(8): 1649-1655. (TAO J H, LU Y M, HE D, et al. Incorporating interaction among user attributes into personalized review summarization[J]. Journal of Chinese computer systems, 2023, 44(8): 1649-1655.)
[27] 慕永利. 基于深度学习的情绪原因识别方法研究[D]. 太原: 山西大学, 2018. (MU Y L. Research on emotional cause detec-tion based on deep learning[D]. Taiyuan: Shanxi University, 2018.)
[28] 胡令传, 陶晓鹏. 客户评论中用户体验信息自动提取研究[J]. 计算机工程, 2015, 41(1): 49-53, 60. (HU L C, TAO X P. Re-search on information automatic extraction of user experience from customer reviews[J]. Computer engineering, 2015, 41(1): 49-53, 60.)
[29] WU Z J, LI L Z, NA C A. Scenario-based user’s needs discovery for creative product design[C]//2009 Second International Symposium on Computational Intelligence and Design.Piscataway: IEEE, 2009: 368-371.
[30] 谷莹, 李贺, 祝琳琳. 融合主题聚类和语义图模型的产品评论自动摘要方法研究[J]. 图书情报工作, 2022, 66(13): 118-126. (GU Y, LI H, ZHU L L. Research on automatic summarization method of product reviews based on topic clustering and se-mantic graph model[J]. Library and information service, 2022, 66(13): 118-126.)
[31] 唐晓波, 刘广超. 细粒度情感分析研究综述[J]. 图书情报工作, 2017, 61(5): 132-140. (TANG X B, LIU G C. Research re-view on fine-grained sentiment analysis[J]. Library and information service, 2017, 61(5): 132-140.)
Options
文章导航

/

〈 〉