Please wait a minute...
Data Analysis and Knowledge Discovery  2023, Vol. 7 Issue (10): 63-73    DOI: 10.11925/infotech.2096-3467.2022.0872
Current Issue | Archive | Adv Search |
Extracting Product Features and Analyzing Customer Needs from Chinese Online Reviews with Hybrid Neural Network
Shi Lili1,2,Lin Jun1,2(),Zhu Guiyang3
1School of Management, Xi’an Jiaotong University, Xi’an 710049, China
2The Key Lab of the Ministry of Education for Process Management & Efficiency Engineering, Xi’an 710049, China
3School of Management, Hangzhou Dianzi University, Hangzhou 310018, China
Download: PDF (1072 KB)   HTML ( 16 ) 
Export: BibTeX | EndNote (RIS)      
Abstract  

[Objective] This study aims to extract product features and analyze customer needs based on the content of Chinese online reviews. [Methods] First, we proposed a hybrid neural network (HNN) to extract product features. Then, we applied critical incident technique (CIT) and analysis of complaints and compliments (ACC) to the Kano model to classify and prioritize product features. [Results] The F1 value of the HNN model reached 94.85%, which was 10.52 percentage points higher than the variant benchmark models and 9.47 percentage points over other leading models on average. [Limitations] The proposed model is supervised learning, and the need for labeling information restricts its application. [Conclusions] The proposed method improves the accuracy of product feature extraction, as well as classifies and prioritizes product features based on customer needs. It lays a foundation for managers to develop product improvement strategies.
Key words： Chinese Online Reviews      Product Feature Extraction      Customer Requirements Analysis      Deep Learning     
Received: 18 August 2022      Published: 28 March 2023
ZTFLH:  TP391  
  F274  
Fund:National Natural Science Foundation of China(72071154);National Natural Science Foundation of China(71672140)
Corresponding Authors: Lin Jun，ORCID：0000-0002-2635-1816，E-mail：ljun@mail.xjtu.edu.cn。   
Service
E-mail this article
Add to my bookshelf
Add to citation manager
E-mail Alert
RSS
Articles by authors
Lili Shi
Jun Lin
Guiyang Zhu

Cite this article:

Shi Lili, Lin Jun, Zhu Guiyang. Extracting Product Features and Analyzing Customer Needs from Chinese Online Reviews with Hybrid Neural Network. Data Analysis and Knowledge Discovery, 2023, 7(10): 63-73.

URL:

https://manu44.magtech.com.cn/Jwk_infotech_wk3/EN/10.11925/infotech.2096-3467.2022.0872     OR     https://manu44.magtech.com.cn/Jwk_infotech_wk3/EN/Y2023/V7/I10/63

The Research Framework
Structure of HNN Algorithm
评论属性 评论长度
产品 Huawei Mate 10 均值 96.14
总数量 139 684 最小值 3
预处理后数量 134 125 最大值 645
网站来源 JD.com 25% 27
评论评分 1~5 50% 53
收集时间 2017.11.1-2018.11.12 75% 118
Overview of the Dataset
模块 参数 值
CNW 词向量维度 300
学习率 0.001
Dropout 0.5
卷积核尺寸 1-3
卷积核数量 每个尺寸各100
激活函数 ReLU
BLC 字向量维度 300
学习率 0.001
Dropout 0.3
BiLSTM个数 10
BiLSTM隐层神经元数量 32
Dropout 0.3
激活函数 ReLU
FC 隐层层数 1
隐层神经元数量 64
Dropout 0.3
激活函数 Softmax
输出维度 13
Hyperparameter Settings
变体 定义
CNW 词向量输入CNN中，CNN输出输入FC中
CNC 字向量输入CNN中，CNN输出输入FC中
BLW 词向量输入BiLSTM中，BiLSTM输出输入FC中
BLC 字向量输入BiLSTM中，BiLSTM输出输入FC中
CNW+CNC 词向量和字向量分别输入两个CNN中，两个CNN输出串联输入FC中
BLW+BLC 词向量和字向量分别输入两个BiLSTM中，两个BiLSTM输出串联输入FC中
BLW+CNC 词向量和字向量分别输入BiLSTM和CNN中，两个输出串联输入FC中
Description of the Variants
编号 变体 P/% R/% F1/%
1 CNW 92.79 94.21 93.49
2 CNC 86.89 94.52 90.54
3 BLW 67.05 72.76 69.79
4 BLC 71.06 76.09 73.49
5 CNW+CNC 90.44 90.09 90.27
6 BLW+BLC 79.83 81.84 80.82
7 BLW+CNC 89.69 94.22 91.90
8 CNW+BLC（HNN） 93.94 95.78 94.85
Performance of the Variants
Performance of the Variants
模型 P/% R/% F1/%
SVM 62.58 95.70 75.67
C-CNN 92.16 95.75 93.92
D-RNN 85.02 88.11 86.54
HNN 93.94 95.78 94.85
Performance of Different Models
产品特征 P/% R/% F1/%
硬件 93.13 92.91 93.02
电池 99.26 94.15 96.64
功能 85.03 96.79 90.53
通信 95.54 96.57 96.05
音视频 97.08 99.11 98.09
性价比 85.63 93.8 89.53
屏幕 96.35 95.19 95.77
质量 82.53 93.49 87.66
相机 99.54 96.93 98.22
系统 96.46 98.01 97.23
外观及体验 94.65 97.68 96.14
包装及配件 96.83 95.05 95.93
物流及售后 99.15 95.48 97.28
Results of Each Product Feature
句子 预测 真实
1.续航时间相当可观 电池 电池
2.默认分辨率显示效果很好 屏幕 屏幕
3.读取和写入原有手机备份很方便 系统 系统
4.我收到的手机WiFi断流 通信 通信
5.稍微有点卡不如苹果流畅 系统 系统
6.店家发货非常快 物流及售后 物流及售后
Examples of Implicit Product Features
产品特征 u i /% v i /% 分类 u i v i
通信 0.508 1.585 必备特征 0.321
硬件 0.729 1.574 必备特征 0.463
音视频 0.859 1.574 必备特征 0.546
性价比 4.207 6.749 必备特征 0.623
物流及售后 25.490 38.434 必备特征 0.663
包装及配件 4.941 7.053 一维特征 0.701
电池 7.121 9.425 一维特征 0.756
屏幕 2.488 3.243 一维特征 0.767
外观及体验 17.734 13.791 一维特征 1.286
功能 2.225 1.627 魅力特征 1.368
质量 0.885 0.514 魅力特征 1.722
相机 9.968 4.775 魅力特征 2.088
系统 22.844 9.656 魅力特征 2.366
Product Feature Classifications Based on Kano
[1] Peng H Y, Ma Y K, Li Y, et al. Learning Multi-Grained Aspect Target Sequence for Chinese Sentiment Analysis[J]. Knowledge-Based Systems, 2018, 148: 167-176.
doi: 10.1016/j.knosys.2018.02.034
[2] Teahan W J, Wen Y Y, McNab R, et al. A Compression-Based Algorithm for Chinese Word Segmentation[J]. Computational Linguistics, 2000, 26(3): 375-393.
doi: 10.1162/089120100561746
[3] 唐琳, 郭崇慧, 陈静锋. 中文分词技术研究综述[J]. 数据分析与知识发现, 2020, 4(2/3): 1-17.
[3] (Tang Lin, Guo Chonghui, Chen Jingfeng. Review of Chinese Word Segmentation Studies[J]. Data Analysis and Knowledge Discovery, 2020, 4(2/3): 1-17.)
[4] Hu M Q, Liu B. Mining Opinion Features in Customer Reviews[C]// Proceedings of the 19th National Conference on Artifical Intelligence. 2004: 755-760.
[5] Qiu G, Liu B, Bu J J, et al. Opinion Word Expansion and Target Extraction Through Double Propagation[J]. Computational Linguistics, 2011, 37(1): 9-27.
doi: 10.1162/coli_a_00034
[6] Blei D M, Ng A Y, Jordan M I. Latent Dirichlet Allocation[J]. The Journal of Machine Learning Research, 2003, 3: 993-1022.
[7] Schouten K, Frasincar F. Survey on Aspect-Level Sentiment Analysis[J]. IEEE Transactions on Knowledge and Data Engineering, 2016, 28(3): 813-830.
doi: 10.1109/TKDE.2015.2485209
[8] Poria S, Cambria E, Gelbukh A. Aspect Extraction for Opinion Mining with a Deep Convolutional Neural Network[J]. Knowledge-Based Systems, 2016, 108: 42-49.
doi: 10.1016/j.knosys.2016.06.009
[9] Kim Y. Convolutional Neural Networks for Sentence Classification[C]// Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing. 2014: 1746-1751.
[10] Liu P F, Joty S, Meng H. Fine-Grained Opinion Mining with Recurrent Neural Networks and Word Embeddings[C]// Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. 2015: 1433-1443.
[11] Li X, Bing L D, Li P J, et al. Aspect Term Extraction with History Attention and Selective Transformation[C]// Proceedings of the 27th International Joint Conference on Artificial Intelligence. 2018: 4194-4200.
[12] Zhou J, Chen Q, Huang J X, et al. Position-Aware Hierarchical Transfer Model for Aspect-Level Sentiment Classification[J]. Information Sciences, 2020, 513: 1-16.
doi: 10.1016/j.ins.2019.11.048
[13] Liu G L, Xu X F, Deng B L, et al. A Hybrid Method for Bilingual Text Sentiment Classification Based on Deep Learning[C]// Proceedings of the 17th IEEE/ACIS International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing. 2016: 93-98.
[14] Zhou K, Long F. Sentiment Analysis of Text Based on CNN and Bi-directional LSTM Model[C]// Proceedings of the 24th IEEE International Conference on Automation and Computing. 2018: 613-617.
[15] Pei J H, Zhang C, Huang D G, et al. Combining Word Embedding and Semantic Lexicon for Chinese Word Similarity Computation[C]// Proceedings of International Conference on Computer Processing of Oriental Languages, National CCF Conference on Natural Language Processing and Chinese Computing. 2016: 766-777.
[16] 杨阳, 刘恩博, 顾春华, 等. 稀疏数据下结合词向量的短文本分类模型研究[J]. 计算机应用研究, 2022, 39(3): 711-715, 750.
[16] (Yang Yang, Liu Enbo, Gu Chunhua, et al. Research on Short Text Classification Model Combined with Word Vector for Sparse Data[J]. Application Research of Computers, 2022, 39(3): 711-715, 750.)
[17] Hashida S, Tamura K, Sakai T. Classifying Sightseeing Tweets Using Convolutional Neural Networks with Multi-Channel Distributed Representation[C]// Proceedings of the 2018 IEEE International Conference on Systems, Man, and Cybernetics. 2018: 178-183.
[18] Chen X X, Xu L, Liu Z Y, et al. Joint Learning of Character and Word Embeddings[C]// Proceedings of the 24th International Conference on Artificial Intelligence. 2015: 1236-1242.
[19] Kano N, Seraku N, Takahashi F, et al. Attractive Quality and Must-Be Quality[J]. Journal of the Japanese Society for Quality Control, 1984, 14(2): 147-156.
[20] Qi J Y, Zhang Z P, Jeon S, et al. Mining Customer Requirements from Online Reviews: A Product Improvement Perspective[J]. Information & Management, 2016, 53(8): 951-963.
doi: 10.1016/j.im.2016.06.002
[21] Mikulić J, Prebežac D. A Critical Review of Techniques for Classifying Quality Attributes in the Kano Model[J]. Managing Service Quality, 2011, 21(1): 46-66.
doi: 10.1108/09604521111100243
[22] Flanagan J C. The Critical Incident Technique[J]. Psychological Bulletin, 1954, 51(4): 327-358.
doi: 10.1037/h0061470 pmid: 13177800
[23] Bott G, Tourish D. The Critical Incident Technique Reappraised: Using Critical Incidents to Illuminate Organizational Practices and Build Theory[J]. Qualitative Research in Organizations and Management, 2016, 11(4): 276-300.
doi: 10.1108/QROM-01-2016-1351
[24] Heo J Y, Kim K J. Development of a Scale to Measure the Quality of Mobile Location-Based Services[J]. Service Business, 2017, 11(1): 141-159.
doi: 10.1007/s11628-016-0305-6
[25] Cadotte E R, Turgeon N. Dissatisfiers and Satisfiers: Suggestions from Consumer Complaints and Compliments[J]. The Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behavior, 1988, 1: 74-79.
[26] Cadotte E R, Turgeon N. Key Factors in Guest Satisfaction[J]. Cornell Hotel and Restaurant Administration Quarterly, 1988, 28(4): 44-51.
doi: 10.1177/001088048802800415
[27] Tontini G, dos Santos Bento G, Milbratz T C, et al. Exploring the Nonlinear Impact of Critical Incidents on Customers’ General Evaluation of Hospitality Services[J]. International Journal of Hospitality Management, 2017, 66: 106-116.
doi: 10.1016/j.ijhm.2017.07.011
[28] Huizing M. Twitter as a Giant Ideabox-Systematically Identifying Customer Needs Regarding the Ring Video Doorbell Through Analysis of Tweets[D]. Enschede, The Netherlands: University of Twente, 2021.
[29] Mikolov T, Chen K, Corrado G, et al. Efficient Estimation of Word Representations in Vector Space[OL]. arXiv Preprint, arXiv: 1301.3781.
[30] Liu W, Xu T G, Xu Q H, et al. An Encoding Strategy Based Word-Character LSTM for Chinese NER[C]// Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics:Human Language Technologies, Volume 1(Long and Short Papers). 2019: 2379-2389.
[31] He R D, Lee W S, Ng H T, et al. Exploiting Document Knowledge for Aspect-Level Sentiment Classification[C]// Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2:Short Papers). 2018: 579-585.
[32] dos Santos C, Gatti M. Deep Convolutional Neural Networks for Sentiment Analysis of Short Texts[C]// Proceedings of the 25th International Conference on Computational Linguistics. 2014: 69-78.
[33] Gu X D, Gu Y W, Wu H B. Cascaded Convolutional Neural Networks for Aspect-Based Opinion Summary[J]. Neural Processing Letters, 2017, 46(2): 581-594.
doi: 10.1007/s11063-017-9605-7
[34] Liu T F, Yu S Y, Xu B M, et al. Recurrent Networks with Attention and Convolutional Networks for Sentence Representation and Classification[J]. Applied Intelligence, 2018, 48(10): 3797-3806.
doi: 10.1007/s10489-018-1176-4
[35] Nowak J, Taspinar A, Scherer R. LSTM Recurrent Neural Networks for Short Text and Sentiment Classification[C]// Proceedings of International Conference on Artificial Intelligence and Soft Computing. 2017: 553-562.
[36] Smith L N. A Disciplined Approach to Neural Network Hyper-Parameters: Part 1 - - Learning Rate, Batch Size, Momentum, and Weight Decay[OL]. arXiv Preprint, arXiv: 1803.09820.
[37] Tamchyna A, Veselovská K. UFAL at SemEval-2016 Task 5: Recurrent Neural Networks for Sentence Classification[C]// Proceedings of the 10th International Workshop on Semantic Evaluation(SemEval-2016). 2016: 367-371.
[1] Xiang Zhuoyuan, Chen Hao, Wang Qian, Li Na. Few-Shot Language Understanding Model for Task-Oriented Dialogues[J]. 数据分析与知识发现, 2023, 7(9): 64-77.
[2] Nie Hui, Cai Ruisheng. Online Doctor Recommendation System with Attention Mechanism[J]. 数据分析与知识发现, 2023, 7(8): 138-148.
[3] Li Guangjian, Yuan Yue. Review of Knowledge Elements Extraction in Scientific Literature Based on Deep Learning[J]. 数据分析与知识发现, 2023, 7(7): 1-17.
[4] Wang Nan, Wang Qi. Evaluating Student Engagement with Deep Learning[J]. 数据分析与知识发现, 2023, 7(6): 123-133.
[5] Wu Jialun, Zhang Ruonan, Kang Wulin, Yuan Puwei. Deep Learning Model of Drug Recommendation Based on Patient Similarity Analysis[J]. 数据分析与知识发现, 2023, 7(6): 148-160.
[6] Wang Xiaofeng, Sun Yujie, Wang Huazhen, Zhang Hengzhang. Construction and Verification of Type-Controllable Question Generation Model Based on Deep Learning and Knowledge Graphs[J]. 数据分析与知识发现, 2023, 7(6): 26-37.
[7] Liu Yang, Zhang Wen, Hu Yi, Mao Jin, Huang Fei. Hotel Stock Prediction Based on Multimodal Deep Learning[J]. 数据分析与知识发现, 2023, 7(5): 21-32.
[8] Huang Xuejian, Ma Tinghuai, Wang Gensheng. Detecting Weibo Rumors Based on Hierarchical Semantic Feature Learning Model[J]. 数据分析与知识发现, 2023, 7(5): 81-91.
[9] Wang Yinqiu, Yu Wei, Chen Junpeng. Automatic Question-Answering in Chinese Medical Q & A Community with Knowledge Graph[J]. 数据分析与知识发现, 2023, 7(3): 97-109.
[10] Zhang Zhengang, Yu Chuanming. Knowledge Graph Completion Model Based on Entity and Relation Fusion[J]. 数据分析与知识发现, 2023, 7(2): 15-25.
[11] Shen Lining, Yang Jiayi, Pei Jiaxuan, Cao Guang, Chen Gongzheng. A Fine-Grained Sentiment Recognition Method Based on OCC Model and Triggering Events[J]. 数据分析与知识发现, 2023, 7(2): 72-85.
[12] Wang Weijun, Ning Zhiyuan, Du Yi, Zhou Yuanchun. Identifying Interdisciplinary Sci-Tech Literature Based on Multi-Label Classification[J]. 数据分析与知识发现, 2023, 7(1): 102-112.
[13] Xiao Yuhan, Lin Huiping. Mining Differentiated Demands with Aspect Word Extraction: Case Study of Smartphone Reviews[J]. 数据分析与知识发现, 2023, 7(1): 63-75.
[14] Cheng Quan, She Dexin. Drug Recommendation Based on Graph Neural Network with Patient Signs and Medication Data[J]. 数据分析与知识发现, 2022, 6(9): 113-124.
[15] Wang Lu, Le Xiaoqiu. Research Progress on Citation Analysis of Scientific Papers[J]. 数据分析与知识发现, 2022, 6(4): 1-15.
  Copyright © 2016 Data Analysis and Knowledge Discovery   Tel/Fax:(010)82626611-6626，82624938   E-mail:jishu@mail.las.ac.cn