随着我国汽车产业的高质量发展与转型，新能源汽车保有量明显陡增，产品质量问题进入集中爆发阶段，严重制约新能源汽车行业发展。产品质量改进对新能源汽车行业的可持续性有着积极影响，新能源汽车用户的年轻化与个性化对产品质量改进的及时性和精准性提出了更高的要求。但受限于传统质量反馈方式（问卷、电话访谈及调研等）无法及时与准确的识别用户质量改进需求。在线评论作为商业情报的可靠来源，为新能源汽车质量改进识别提供了新的生态土壤。

本文基于在线评论质量反馈视角，运用深度学习、依存句法分析、Kano模型和质量功能展开QFD等方法，构建了新能源汽车产品质量改进识别框架，并针对新能源汽车T插电混动版进行了实证研究，具体完成以下工作。

（1）首先完善了新能源汽车质量改进微观理论。从汽车与其他产品、新能源汽车与传统汽车质量改进异同进行微观分析，为新能源汽车产品质量改进建立基础。（2）其次构建了基于在线评论的新能源汽车用户质量改进识别框架。抓取35种热销新能源汽车评论，对评论进行分词、停用词去除和词性标注等预处理。为准确识别质量问题选择长短期记忆模型LSTM分类负面观点评论，对所得负面评论进行依存句法分析得到6条方面级观点抽取规则，结合词典对方面级观点进行量化，并采用Kano模型进行需求分类。（3）最后构建了基于在线评论的质量功能展开QFD方法，针对新能源汽车T插电式混动版产品进行了实证研究。识别出15个用户质量改进需求，运用质量工具亲和图KJ对质量改进需求进行层次分解，邀请专家分析与打分得到17条相对应的工程质量特性和关系矩阵，然后利用用户评分得到市场竞争矩阵。将所有矩阵输入质量屋，得出产品质量改进相关结论与建议，进一步证明了方法的可行性与有效性。

本文提出基于在线评论的新能源汽车质量问题识别框架，为在线评论驱动质量改进识别提供思路、框架和工具方法，补充和深化了大数据驱动的质量管理研究。不仅有助于企业识及时别新能源汽车质量隐患，亦可辅助相关质量与设计人员更好地了解用户质量改进需求，为产品质量改进提供决策参考。

With the high-quality development and transformation of China's automobile industry, the number of new energy vehicles has increased significantly, and the product quality problems have entered a concentrated outbreak stage, which seriously restricts the development of new energy vehicle industry. Product quality improvement has a positive impact on the sustainability of the new energy vehicle industry. The younger and personalized new energy vehicle users put forward higher requirements on the timeliness and accuracy of product quality improvement. However, due to the traditional quality feedback methods (questionnaire, telephone interview and research, etc.), it is unable to timely and accurately identify users' quality improvement needs. As a reliable source of business intelligence, online reviews provide a new ecological soil for the identification of new energy vehicle quality improvement.

Based on the quality feedback perspective of online reviews, this paper uses deep learning, dependency parsing, Kano model and quality function deployment QFD methods to build a new energy vehicle product quality improvement identification framework, and carries out an empirical study on the T plug-in hybrid version of new energy vehicles, and specifically completes the following work.

(1) Firstly, the micro theory of new energy vehicle quality improvement is improved. This paper analyzes the similarities and differences of quality improvement between automobile and other products, new energy vehicle and traditional vehicle, and establishes a theoretical basis for the quality improvement of new energy vehicle products. (2) Secondly, a new energy vehicle user quality improvement identification framework based on online reviews is constructed. 35 kinds of hot new energy vehicle reviews are captured and preprocessed by word segmentation, stop word removal and part of speech tagging. In order to identify quality problems accurately, the short-term and long-term memory model LSTM is selected to classify negative comments. The negative comments are analyzed by dependency syntax, and six aspect level opinion extraction rules are obtained. Combined with the dictionary, the aspect level opinions are quantified, and the Kano model is used for demand classification. (3) Finally, the QFD method based on online reviews is constructed, and an empirical study is carried out on the T plug-in hybrid version of new energy vehicles. 15 users' quality improvement requirements are identified, and the quality tool affinity graph kJ is used to decompose the quality improvement requirements hierarchically. Experts are invited to analyze and score 17 corresponding engineering quality characteristics and relationship matrices, and then the market competition matrix is obtained by using the user score. All the matrices are input into the house of quality, and the conclusions and suggestions related to product quality improvement are obtained, which further proves the feasibility and effectiveness of the method.

This paper proposes a new energy vehicle quality problem identification framework based on online review, which provides ideas, framework and tools for solving quality improvement problems driven by review data, and supplements and deepens the research on quality management driven by big data. It not only helps enterprises to identify new energy vehicle quality risks, but also helps relevant quality and design personnel to better understand the quality improvement needs of users, and provides decision-making reference for product quality improvement.

[1]2020年全国消协组织受理投诉情况分析[EB/OL].中国消费者协会,2021-02-03.http://www.cca.org.cn/tsdh/detail/29923.html

[2]2020年12月汽车工业经济运行情况[EB/OL].中华人民共和国工业和信息化,2021-01-14.https://www.miit.gov.cn/gxsj/tjfx/zbgy/qc/art/2021/art_80a68716879e4180a0dea39183b546cf.html

[3]2020年12月召回公告[EB/OL].国家市场监督管理总局缺陷产品管理中心,2020-12-31.http://dpac.samr.gov.cn/

[4]任明仑,宋月丽.大数据:数据驱动的过程质量控制与改进新视角[J].计算机集成制造系统,2019,25(11):2731-2742.

[5]杨善林,周开乐.大数据中的管理问题:基于大数据的资源观[J].管理科学学报,2015,18(05):1-8.

[6]刘书庆,熊琪凯.产业高质量发展对产品实现质量影响实证研究[J/OL].工业工程与管理:1-10[2021-03-04].http://kns.cnki.net/kcms/detail/31.1738.T.20201124.1531.004.html.

[7]张瑞,金志刚,胡博宏,张子洋.一种情感分析与质量控制的异常评论识别方法[J].哈尔滨工业大学学报,2018,50(09):164-170.

[8]李少波,全华凤,胡建军,吴永明,张安思.基于在线评论数据驱动的产品感性评价方法[J].计算机集成制造系统,2018,24(03):752-762.

[9][17]孙淑慧,朱立龙.消费者反馈下企业质量改进投入决策Moran演化分析[J/OL].中国管理科学:1-12[2021-03-04].https://doi.org/10.16381/j.cnki.issn1003-207x.2019.0678.

[9]齐红倩,李民强,王智鹏.相对质量的现实构造——基于需求因素的经济学分析[J].经济管理,2010,32(06):172-177.

[10]Dick J, Hull E, Jackson K. Requirements engineering[M]. Fourth Edition. London: Springer, 2017.

[11]Boehm B, Basili V R. Software defect reduction top 10 list[J]. Computer, 2001, 34(1):135-137.

[12]李家锋,刘智华.质量管理综述[J].华南农业大学学报,1994(01):133-138.

[13]中华人民共和国国家质量监督检验检疫总局.GB/T19000-2008质量管理体系基础和术语[S].中国标准出版社,2009,(2):4-6.

[14]吴双,刘远,郝晶晶.复杂产品供应链质量改进成本分担契约研究[J].工业工程与管理,2019,24(05):56-63+71.

[15]韩小鹏,张旭梅,伏红勇,张杨.制造商产品改进与增值性服务的选择决策研究[J].管理学报,2016,13(07):1081-1089.

[16]李佩,魏航.基于企业技术水平的改进型新产品最优上市时间研究[J].管理工程学报,2019,33(01):144-158.

[18][49]张流洋,李荷皎,海本禄,李胜坤.基于产品质量改进的动态多响应稳健性集成建模策略[J].统计与决策,2020,36(21):176-180.

[19]Kunkel S, Rosenqvist U, Westerling R. Quality improvement designs are related to the degree of organisation of quality systems: An empirical study of hospital departments[J] Health Policy, 2007, 84(40:191-199.

[20]Agus A, Hassan Z. Enhancing Production Performance and Customer Performance Through Total Quality Management (TQM): Strategies For Competitive Advantage [J]. Procedia Social and Behavioral Sciences, 2011. 24(9):1650-1662.

[21]Lai X, Xie M, Tan K C, Yang B. Ranking of customer requirements in a competitive environment [J]. Computers&Industrial Engineering, 2008, 54(7):202-214.

[22]刘书庆,董丽娜.质量改进有效性对产品实现过程影响实证研究[J].工业工程与管理,2014,19(06):12-21.

[23]刘航.基于知识视角改进的复杂产品系统创新过程研究[J].科学管理研究,2012,30(05):45-47.

[24]张璐,吴菲菲,黄鲁成.基于用户网络评论信息的产品创新研究[J].软科学,2015,29(05):12-16.

[25]Song W. requirement management for product-service systems: Status review and future trends[J]. Computers in Industry, 2017, 85:11-22.

[26]车阿大,林志航,方勇.模糊集理论在QFD中的应用[J].系统工程理论方法应用,1998(02):57-59+64.

[27]王昊琪,李浩,文笑雨.基于本体的公理化系统设计语义建模与推理规则[J/OL].机械工程学报:1-16[2021-03-07].http://kns.cnki.net/kcms/detail/11.2187.TH.20200324.1513.024.html.

[28]苏珂,崔元.面向相似认知用户集群的Triz超系统资源需求获取模型[J/OL].计算机集成制造系统:1-20[2021-03-07].http://kns.cnki.net/kcms/detail/11.5946.tp.20200415.1846. 002.html.

[29]车阿大,林志航.质量功能配置的多目标规划模型[J].计算机集成制造系统-CIMS,1998(06):27-31.

[30]张海沸.QFD在汽车内饰件开发方面的应用[J].上海交通大学学报,2007(S1):180-187.

[31]王晓暾,熊伟.基于QFD和TRIZ的可信软件技术冲突解决方法[J].航空学报,2011,32(01):128-136.

[32]潘振华,刘海江.面向个性化车身产品客户动态需求获取与预测的扩展QFD研究[J].机械科学与技术,2014,33(04):564-572.

[33]万延见,饶宾期,梁喜凤,卢锡龙.一种驱动市场导向式产品创新设计方法[J].计算机集成制造系统,2017,23(07):1369-1376.

[34]吴小丽,熊会元,于丽敏.基于QFD电动汽车技术特性重要度的确定方法研究[J].机械设计与制造,2015(07):131-134+138.

[35]张雷,钟言久,袁远,李璟,秦旭,董万富.基于数据挖掘的绿色设计中客户需求向工程特性权重转化方法[J].中国机械工程,2019,30(02):174-182.

[36]聂大安,李彦,麻广林,马涛.基于用户需求分类的同步多产品设计方法[J].计算机集成制造系统,2010(06):1131-1137.

[37]尹蕾,蒋建国,路瑞刚,尹安东.面向大数据的新能源汽车复杂联盟云评价研究[J].汽车工程,2019,41(01):112-119.

[38]Robert G C, Kogan Page. Winning at new products[J]. Pergamon, 1989, 22(5): 269.

[39]Dyer J H. Effective Inter Firm Collaboration:How Firms Minimize transaction Costs and Maximize Transaction Value [J]. Strategic Management Journal, 1997, 8(2):535-556.

[40]朱志刚,张忠能,凌君逸.APQP质量模块的研究和设计[J].计算机工程,2004(S1): 463-466.

[41]Sungjoo L , Chanwoo C , Jaehong C , et al. R&D Project Selection Incorporating Customer-Perceived Value and Technology Potential: The Case of the Automobile Industry[J]. Sustainability, 2017, 9(10):1918-1918.

[42]段心林,陈科,李天博.汽车研发组织精益质量管理体系构建及优化方法研究[J].科技管理研究,2019,39(10):217-222.

[43]Zuo W , Zhu W , Chen S , et al. Service quality management of online car-hailing based on PCN in the sharing economy[J]. Electronic Commerce Research and Applications, 2019.

[44]You X , Ma J , Zhang Y , et al. VDRF: Sensing the defect information to risk level of vehicle recall based on bert communication model[J]. Computer Science and Information Systems, 2020, 17(3):795-817.

[45]王鸿鹭,蒋炜,魏来,黄文坡.基于物联网的产品全生命周期质量管理的模式创新与展望[J/OL].系统工程理论与实践:1-13[2021-03-09].http://kns.cnki.net/kcms/detail/11.2267.n.20201110.1731.004.html.

[46]Ch F A, Khobreh M, Nasiri S, et al. Knowledge Management Support for Quality Management to Achieve Higher Customer Satisfaction[C]. Electro/Information Technology, 2009. eit '09. IEEE International Conference on. IEEE, 2009.

[47]Weckenmann A, Akkasoglu G, Wemer T. Quality management-history and trends[J]. TQM Journal 2015, 27(3): 281-293.

[48]余厚云,王慧青,张辉,胡玉坤.产品表面质量视觉检测中的相机位姿自动校准[J].东南大学学报(自然科学版),2020,50(05):942-949.

[50]谷莹,李贺,李叶叶,刘嘉宇.基于在线评论的企业竞争情报需求挖掘研究[J].现代情报,2021,41(01):24-31.

[51]Jin Jian, Ji Ping, Gu Rui. Identifying comparative customer requirements from product online reviews for competitor analysis[J]. Engineering Applications of Artificial Intelligence, 2016, 49(C):61-73.

[52]王楠,王莉雅,李瑶,陈劲.同侪影响对用户贡献行为的作用研究——基于网络客观大数据的分析[J/OL].科学学研究:1-18[2021-03-10].https://doi.org/10.16192/j.cnki.1003-2053.20210201.002.

[53]于兆吉,贾宝禹,赵英姿.OTA在线信誉系统对消费者购买决策的影响研究[J].中国软科学,2021(01):147-155.

[54]倪峰,李永明,郑德俊,沈军威.移动图书馆服务平台的改进需求识别[J].图书情报工作,2016,60(21):17-23.

[55]宋严.社交媒体文本信息多层次细粒度属性挖掘方法研究[J].情报科学,2020,38(11):98-103.

[56]Wilks, Ronald. Short stories[M]. Penguin, 1984.

[57]Dave K, Lawrence S, Pennock D M. Mining the peanut gallery: Opinion Extraction and Semantic Classification of Product Reviews[J]. 2003, 289(51): 35326-40.

[58]Liu B. Sentiment Analysis and Opinion Mining[C]. Synthesis Lectures on Human Language Technologies 5.1 (2012): 1-167. Morgan&Claypool, 2011.

[59]徐久成,沈钧毅,安秋生,李乃乾.基于信息粒度与粗糙集的决策细化研究[J].西安交通大学学报,2005(04):335-338.

[60]王俊,史存会,张瑾,俞晓明,刘悦,程学旗.融合上下文信息的篇章级事件时序关系抽取方法[J/OL].计算机研究与发展:1-9[2021-03-10].http://kns.cnki.net/kcms/detail/11.1777.TP.20210203.1208.006.html.

[61]韩忠明,李梦琪,刘雯,张梦玫,段大高,于重重.网络评论方面级观点挖掘方法研究综述[J].软件学报,2018,29(02):417-441.

[62]于超,樊治平.基于顾客在线评论的服务要素优化配置方法[J].计算机集成制造系统,2019,25(03):714-725.

[63]张严,李天瑞.面向评论的方面级情感分析综述[J].计算机科学,2020,47(06):194-200.

[64]曾锋,曾碧卿,韩旭丽,张敏,商齐.基于双层注意力循环神经网络的方面级情感分析[J].中文信息学报,2019,33(06):108-115.

[65]吕品,钟珞,蔡敦波,吴云韬.基于CRF的中文评论有效性挖掘产品特征[J].计算机工程与科学,2014,36(02):359-366.

[66]Pang B, Lee L. Opinion Mining and Sentiment Analysis[J]. Foundations and Trends? In Information Retrieval, 2008, 2(1–2): 1-135.

[67]陈虹,杨燕,杜圣东.用户评论方面级情感分析研究[J].计算机科学与探索,2021,15(03):478-485.

[68]汤凌燕,熊聪聪,王嫄,周宇博,赵子健.基于深度学习的短文本情感倾向分析综述[J/OL].计算机科学与探索:1-18[2021-03-11].http://kns.cnki.net/kcms/detail/11.5602.TP.20210204.1141.004.html.

[69]丁亚龙,谌云莉.基于记忆增强和知识迁移的方面级用户评论情感分析[J].计算机应用研究,2020,37(S2):31-33.

[70]秦春秀,祝婷,赵捧未,张毅.自然语言语义分析研究进展[J].图书情报工作,2014,58(22):130-137.

[71]Garey H B, Tesnière, Lucien, Tesniere L. Esquisse dune syntaxe structurale[J]. Language, 1954, 30(4):512.

[72]李纲,刘广兴,毛进,叶光辉.一种基于句法分析的情感标签抽取方法[J].图书情报工作,2014,58(14):12-20.

[73]喻影,陈珂,寿黎但,陈刚,吴晓凡.基于关键词和关键句抽取的用户评论情感分析[J].计算机科学, 2019,46(10):19-26.

[74]周知,方正东.融合依存句法与产品特征库的用户观点识别研究[J/OL].情报理论与实践:1-14[2021-03-09]. http: //kns. cnki. net/kcms/detail/11. 1762. G3. 20210208. 1513. 015. html.

[75]兰秋军,刘文星,李卫康,胡星野.融合句法信息的金融论坛文本情感计算研究[J].现代图书情报技术,2016(04):64-71.

[76]熊伟.质量功能展开概念的扩展与在软件中的应用[J].管理工程学报,1999(03):71-72.

[77]Han Z M, Li M Q, Liu W, et al. Survey of studies on aspect based opinion mining of Internet[J]. Journal of Software, 2018, 29(2):417-441.

[78]康丹,赵福全,刘宗巍.汽车与智能手机产品开发过程对比分析与改进策略研究[J].科技管理研究,2018,38(02):112-118.

[79]佘承其,张照生,刘鹏,孙逢春.大数据分析技术在新能源汽车行业的应用综述——基于新能源汽车运行大数据[J].机械工程学报,2019,55(20):3-16.

[80]颜端武,杨雄飞,李铁军.基于产品特征树和LSTM模型的产品评论情感分析[J].情报理论与实践,2019,42(12):134-138.

[81]邓楠. 基于LSTM的汽车评论文本分类研究与应用[D].合肥工业大学,2018.

[82]Che Wanxiang, Li Zhenghua, Liu Ting LTP:a Chinese language technology platform[C].The 23rd International Conference on Computational Linguistics. Beijing: TSinghua University press, 2010:3-16.

[83]常建鹏,陈振颂,王先甲,张军.不确定环境下基于质量功能展开的电动汽车技术特性综合重要度确定[J].计算机集成制造系统,2020,26(01):103-113.