| 443 | 8 | 494 |
| 下载次数 | 被引频次 | 阅读次数 |
不少基于智能网联汽车（ICV）的研究均假设用户高度接受并同质使用智能座舱内的智能网联功能，而在汽车“新四化”的不同阶段，用户对ICV的功能需求及使用行为有着显著个体差别特征。忽略这些特征，对智能汽车相关研究及技术发展均有较大影响。本文基于感性工学分析方法及利用大量用户的评论文本大数据，提出群体感性工学理念，旨在建立一套系统科学的ICV功能需求与偏好量化分析体系，为智能网联背景下的驾驶员行为研究及ICV技术发展提供决策支撑。该技术框架包括高效的文本数据挖掘分析方法，所提出的基于汽车领域增量训练的Bert模型能自动化识别用户评论文本中的意图、目标功能及情感强度。最后，通过引入座舱内用户行为数据建立NLP-Kansei模型，能大幅提升模型的预测性能。通过对某新能源汽车品牌两款平台车型85 441名车主共138 819条用户评论进行模型应用，精细化挖掘不同人群对于智能网联功能偏好的细微差别。对比传统预训练模型，新模型能有效挖掘用户需求偏好并显著提升计算效率，其技术框架及结果能为相关领域研发提供有效支撑。
Abstract：In recent research on intelligent connected vehicles(ICVs), it was assumed that drivers would homogeneously accept and use all ICV functions in the smart cockpit, which is not true i.e., individual preferences were not considered. Omitting differences in ICV preference may lead to significant shortcomings of ICV-related research and technology development. This paper proposes the concept of collective-intelligence Kansei engineering with a large number of user-generated comments for developing a systematic evaluation method for quantifying the functionality requirements and preferences. The proposed technology framework includes an efficient text-mining algorithm, and an incremental pretrained Bert model is employed that can automatically identify user purposes, targeted functions, and emotion polarity from user texts. Finally, an NLP-Kansei model is developed with user behaviors from the smart cockpit, and the incorporation of the Bert model and user activities can significantly improve the performance of the classification targets. Data of 85441 drivers and 138819user comments from two new energy vehicle platforms were used for testing the model. The results indicated that the proposed model outperformed the conventional Bert model. The proposed technology framework is practical and can effectively support ICV-related research and development.
[1]张颖.新能源车是2021年车市最大亮点[J].汽车与配件, 2022(2):4.ZHANG Ying. New energy vehicles are the biggest highlight of the auto market in 2021[J]. Automobile&Parts,2022(2):4.
[2]祁宏生,应雨燕,林俊山,等.混合自动驾驶场景多换道需求下的主动间隙适配和换道序列规划[J].交通运输工程与信息学报, 2021, 19(4):36-51.QI Hong-sheng, YING Yan-yu, LIN Jun-shan, et al. Proactive gap adaption and sequence planning for multiple lane-changing requests under mixed autonomous vehicle flow[J]. Journal of Transportation Engineering and Information, 2021, 19(4):36-51.
[3]姚志洪,郝慧君,巫雪梅,等.考虑自动驾驶的混合交通流路段阻抗函数[J].交通运输工程与信息学报, 2021,19(4):1-12.YAO Zhi-hong, HAO Hui-jun, WU Xue-mei, et al. Cost function of mixed traffic flow with autonomous driving[J]. Journal of Transportation Engineering and Information, 2021, 19(4):1-12.
[4]沈悦,陈璟,周子涵,等.车联网环境下自动驾驶车辆动态障碍物协作避让模型[J].交通运输工程与信息学报,2021, 19(4):13-23.SHEN Yue, CHEN Jing, ZHOU Zi-han, et al. Model for cooperative dynamic obstacle avoidance of automated vehicle swarms in connected vehicles environments[J]. Journal of Transportation Engineering and Information, 2021,19(4):13-23.
[5]孙超,黄愉文,张凯,等.智能网联汽车产业政策趋势分析及发展思考[J].城市交通, 2022, 20(1):52-58.SUN Chao, HUANG Yu-wen, ZHANG Kai, et al. Policy analysis and development of connected and autonomous vehicles industry[J]. Urban Transport of China, 2022, 20(1):52-58.
[6]谭征宇,戴宁一,张瑞佛,等.智能网联汽车人机交互研究现状及展望[J].计算机集成制造系统, 2020, 26(10):2615-2632.TAN Zheng-yu, DAI Ning-yi, ZHANG Rui-fo, et al. Overview and perspectives on human-computer interaction in intelligent and connected vehicles[J]. Computer Integrated Manufacturing Systems, 2020, 26(10):2615-2632.
[7]张晓聪.汽车智能座舱发展现状及未来趋势[J].汽车纵横, 2019(8):42-45.ZHANG Xiao-cong. Development status and future trends of automotive smart cockpit[J]. Auto Review, 2019(8):42-45.
[8] LAI X, ZHANG S, MAO N, et al. Kansei engineering for new energy vehicle exterior design:anInternet big data mining approach[J]. Computers&Industrial Engineering,2022, 165:107913.
[9] LAI X, ZHANG Q, CHEN Q, et al. The analytics of product-design requirements using dynamic internet data:application to Chinese smartphone market[J]. International Journal of Production Research, 2019, 57(18):5660-5684.
[10]李贺,谷莹,刘嘉宇.数据驱动下基于语义相似性的产品需求识别研究[J].情报理论与实践, 2022, 45(5):99-106.LI He, GU Ying, LIU Jia-yu. Research on data driven product demand recognition based on semantic similarity[J]. Information Studies:Theory&Application, 2022, 45(5):99-106.
[11]苗红,王艳,黄鲁成,等.基于结构主题模型的用户功能需求挖掘研究[J].情报杂志, 2020, 39(7):134-141.MIAO Hong, WANG Yan, HUANG Lu-cheng, et al. Research on user functional requirements mining based on structured topic model[J]. Journal of Intelligence, 2020,39(7):134-141.
[12]曹阳.基于在线评论数据挖掘的用户需求研究[D].长春:吉林大学, 2020.CAO Yang. User demand research based on online reviews data mining[D]. Changchun:Jilin University, 2020.
[13] CAO H, LIN M. Mining smartphone data for app usage prediction and recommendations:a survey[J]. Pervasive and Mobile Computing, 2017, 37:1-22.
[14] GAO C, XU H, HU J, et al. AR-tracker:track the dynamics of mobile apps via user review mining[C]//2015IEEE Symposium on Service-Oriented System Engineering. San Francisco, USA:IEEE, 2015:284-290.
[15] IACOB C, HARRISON R. Retrieving and analyzing mobile apps feature requests from online reviews[C]//201310th Working Conference on Mining Software Repositories(MSR). San Francisco, USA:IEEE, 2013:41-44.
[16]王伟,王洪伟,孟园.协同过滤推荐算法研究:考虑在线评论情感倾向[J].系统工程理论与实践, 2014, 34(12):3238-3249.WANG Wei, WANG Hong-wei, MENG Yuan. The collaborative filtering recommendation based on sentiment analysis of online reviews[J]. Systems Engineering-Theory&Practice, 2014, 34(12):3238-3249.
[17] ANAM A I, YEASIN M. Accessibility in smartphone applications:what do we learn from reviews?[C]//Proceedings of the 15th International ACM SIGACCESS Conference on Computers and Accessibility. Bellevue, New York:ACM, 2013:1-2.
[18] GUZMAN E, MAALEJ W. How do users like this feature? A fine grained sentiment analysis of app reviews[C]//2014 IEEE 22nd International Requirements Engineering Conference(RE). Karlskrona, Sweden:IEEE,2014:153-162.
[19] HINTON G E, OSINDERO S, TEH Y W. A fast learning algorithm for deep belief nets[J]. Neural Computation,2006, 18(7):1527-1554.
[20] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6):84-90.
[21] KIM Y. Convolutional neural networks for sentence classification[EB/OL].(2014-09-13)[2022-12-20]https://arxiv. org/abs/1408.5882.
[22] Michael I. Jordan. Chapter 25 Serial order:a parallel distributed processing approach[J]. Advances in Psychology, 1997, 121:471-495.
[23] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8):1735-1780.
[24] GRAVES A, SCHMIDHUBER J. Framewise phoneme classification with bidirectional LSTM and other neural network architectures[J]. Neural Networks, 2005, 18(5/6):602-610.
[25] MIKOLOV T, CHEN KAI, CORRADO G, et al. Efficient estimation of word representations in vector space[EB/OL].(2013-09-07)[2022-12-20] https://arxiv. org/abs/1301.3781.
[26] BAHDANAU D, CHO K, BENGIO Y. Neural machine translation by jointly learning to align and translate[EB/OL].(2016-05-19)[2022-12-20] https://arxiv. org/abs/1409.0473.
[27]吴小华,陈莉,魏甜甜,等.基于Self-Attention和BiLSTM的中文短文本情感分析[J].中文信息学报,2019, 33(6):100-107.WU Xiao-hua, CHEN Li, WEI Tian-tian, et al. Sentiment analysis of Chinese short text based on self-attention and Bi-LSTM[J]. Journal of Chinese Information Processing, 2019, 33(6):100-107.
[28]方英兰,孙吉祥,韩兵.基于BERT的文本情感分析方法的研究[J].信息技术与信息化, 2020(2):108-111.FANG Ying-lan, SUN Ji-xiang, HAN Bing. Research on text emotion analysis method based on BERT[J]. Information Technology and Informatization, 2020(2):108-111.
[29]王立梅,朱旭光,汪德嘉,等.基于深度学习的民事案件判决结果分类方法研究[J].计算机科学, 2021, 48(8):80-85.WANG Li-mei, ZHU Xu-guang, WANG De-jia, et al.Study on judicial data classification method based on natural language processing technologies[J]. Computer Science, 2021, 48(8):80-85.
[30]林佳瑞,程志刚,韩宇,等.基于BERT预训练模型的灾害推文分类方法[J].图学学报, 2022, 43(3):530-536.LIN Jia-rui, CHENG Zhi-gang, HAN Yu, et al. Disaster tweets classification method based on pretrained BERT model[J]. Journal of Graphics, 2022, 43(3):530-536.
[31] MINAEE S, KALCHBRENNER N, CAMBRIA E, et al.Deep learning:based text classification:a comprehensive review[J]. ACM Computing Surveys, 2022, 54(3):1-40.
[32] QIU X, SUN T, XU Y, et al. Pre-trained models for natural language processing:a survey[J]. Science China Technological Sciences,2022,63:1872-1897.
[33] YAO X, ZHENG Y. NLP from scratch without largescale pretraining:a simple and efficient framework[EB/OL].(2022-07-22)[2022-12-21] https://arxiv. org/abs/2111.04130.
[34] DEVLIN J, CHANG M W, LEE K, et al. BERT:Pretraining of Deep Bidirectional Transformers for Language Understanding[EB/OL].(2019-05-24)[2022-12-20] https://arxiv.org/abs/2019.1810.04805
[35] GURURANGAN S, MARASOVI?A, SWAYAMDIPTA S, et al. Don't stop pretraining:adapt language models to domains and tasks[EB/OL].(2020-05-05)[2022-12-20] https://arxiv.org/abs/2004.10964.
[36] HUANG Y, DU C, XUE Z, et al. What makes multimodal learning better than single(provably)[EB/OL].(2021-10-26)[2021-6-20] https://arxiv. org/abs/2106.04538.
[37] LIN S, ZOU J, ZHANG C, et al. Understanding user requirements for smart cockpit of new energy vehicles:a natural language process approach[C]//SAE Technical Paper Series. 400 Commonwealth Drive, Warrendale,United States:SAE International, 202-01-7075.
[38] ZHANG C, LAI X, ZOU J, et al. The timing of change:decision making for product optimization with dynamic Kansei engineering[C]//2022 International Conference on Frontiers of Communications, Information System and Data Science(CISDS). Guangzhou, China:IEEE, 2023:1-7.
[39] ZHANG C, ZOU J, LIN S, et al. Mining, monitoring and capturing users’preferences for new energy automobiles with time series analysis on texts sentiments[C]//The 2022 POMS International Conference in China,Xi’an, 2022-06-24~06-27.
基本信息:
DOI：10.19961/j.cnki.1672-4747.2023.01.005
中图分类号:U463.6
引用信息:
[1]赖信君,林深和,邹靖凯,等.基于群体感性工学的智能网联汽车功能偏好分析[J],2023,21(03):59-73.DOI:10.19961/j.cnki.1672-4747.2023.01.005.
基金信息:
广东省自然科学基金资助项目（2114050003127）
2023-01-09
2023
2023-05-26
2023
1