产品评论文本中特征词提取及其关联模型构建与应用

中国机械工程

产品评论文本中特征词提取及其关联模型构建与应用

余琦玮1;肖颖1;林静1;徐新胜1;王庆林1;张飞2

1.中国计量大学工业工程研究所，杭州,310018
2.中国计量大学机械设计制造及其自动化研究所，杭州,310018

出版日期:2017-11-25 发布日期:2017-11-23
基金资助:
国家自然科学基金资助项目（51405462,51305417）;
浙江省自然科学基金资助项目（LY16G010006）;
浙江省科技厅公益性技术应用研究计划资助项目（2014C31117）
National Natural Science Foundation of China （No. 51405462,51305417）
Zhejiang Provincial Natural Science Foundation of China （No. LY16G010006）

Feature Extraction and Correlation Model Construction of Online Product Reviews and Its Applications

YU Qiwei1;IAO Ying1;IN Jing1;U Xinsheng1;ANG Qinglin1;HANG Fei2

1.Institute of Industrial Engineering,China Jiliang University,Hangzhou，310018
2.Institute of Mechanical Design, Manufacturing and Its Automation,China Jiliang University,Hangzhou，310018

Online:2017-11-25 Published:2017-11-23
Supported by:
National Natural Science Foundation of China （No. 51405462,51305417）
Zhejiang Provincial Natural Science Foundation of China （No. LY16G010006）

摘要/Abstract

摘要： 网络上产品评论文本是用户对产品的评价与反馈，及时、有效挖掘其中有价值的信息是制造企业、销售商获取竞争优势迫切需要解决的问题。综合词形、词性、依存关系、控制词及其情感描述等，设计了特征词提取规则单元以及规则模板，基于条件随机场实现了产品特征词的有效提取，并对特征词进行分类；构建了特征词频次、情感评分的计算模型；结合产品特征词的内容与分类，构建了产品特征词关联模型。在此基础上，提出了基于贝叶斯网络的产品特征词关键影响因素推理方法，并以某手机产品为对象进行应用与验证。研究结果可以为制造企业、销售商的精细化管理提供实施依据。

关键词: 文本挖掘, 特征词提取, 情感评分, 关联模型构建, 影响因素推理

Abstract: Online product reviews were the feedback of customer valuing a product. It was an urgent problem for manufacturers and retailers to mine valuable informations effectively and timely from online product reviews with the goal of gaining competitive advantages. Considering comprehensive factors such as word, part-of-speech (POS), dependency relations, governing word and its opinion description, the unit of rule for extracting product features and the rule template were designed. Product features were extracted from online reviews effectively through conditional random field (CRF) theory, and the product features were classified. The quantitative calculation models of product features including frequency and sentiment score were proposed. A correlation model among product features was established based on the description contents of product features and their classifications. On the basis of these, an approach of inferring the key influence factors among product features was presented based on Bayes network. Finally, a case study was performed to verify the feasibility of the methods mentioned above by using a mobile phone as an example, and the results may be used as evidence to implement precision management for manufacturers and retailers.

Key words: text mining, feature extraction, emotional scoring, correlation model construction, influence factor deducing

中图分类号:

TP14
TH128

余琦玮1;肖颖1;林静1;徐新胜1;王庆林1;张飞2. 产品评论文本中特征词提取及其关联模型构建与应用[J]. 中国机械工程.

YU Qiwei1;IAO Ying1;IN Jing1;U Xinsheng1;ANG Qinglin1;HANG Fei2. Feature Extraction and Correlation Model Construction of Online Product Reviews and Its Applications[J]. China Mechanical Engineering.

参考文献

[1]王永，张勤，杨晓洁. 中文网络评论中产品特征提取方法研究[J].情报分析与研究，2013，12:70-73.
WANG Yong, ZHANG Qin, YANG Xiaojie. Research on the Method of Extracting Features from Chinese Product Reviews on the Internet [J]. Intelligence Analysis and Research， 2013, 12: 70-73.
[2]JIN J, JI P, GU R. Identifying Comparative Customer Requirements from Product Online Reviews for Competitor Analysis [J]. Engineering Applications of Artificial Intelligence, 2016,49: 61-73.
[3]RAVI K V , RAGHUVEER K. Web User Opinion Analysis for Product Features Extraction and Opinion Summarization [J]. International Journal of Web & Semantic Technology, 2012,3(4): 69-82.
[4]HU M, LIU B. Mining Opinion Features in Customer Reviews [C]// National Conference on Artificial Intelligence. San Jose, 2004:755-760.
[5]WEI C P, CHEN Y M, YANG C S, et al. Understanding What Concerns Consumers: a Semantic Approach to Product Feature Extraction from Consumer Reviews [J]. Information System & E-business Management, 2010,8(2):149-167.
[6]余传明，陈雷，张小青.基于支持向量机的产品属性识别研究[J].情报学报，2010,29（6）：1038-1044.
YU Chuanming, CHEN Lei, ZHANG Xiaoqing. Mining Product Features from Free-text Customer Reviews: an SVM-based Approach[J]. Journal of the China Society for Scientific & Technical Information, 2010,29（6）：1038-1044.
[7]JAKOB N, GUREVYCH I. Extracting Opinion Targets in a Single- and Cross- Domain Setting with Conditional Random Fields [C]// Conference on Empirical Methods in Natural Language Processing. Cambridge,2010:1035-1045.
[8]SU S H, LYNN K T. Extracting Product Features and Opinion Words Using Pattern Knowledge in Customer Reviews [J]. The Scientific World Journal, 2013: 394758.
[9]李实,叶强,李一军, 等. 中文网络客户评论的产品特征挖掘方法研究[J].管理科学学报， 2009,12(2):142-152.
LI Shi, YE Qiang, LI Yijun, et al. Mining Features of Products from Chinese Customer Online Reviews [J]. Journal of Management Science in China, 2009,12(2):142-152.
[10]LI Shi, YU Ming. Mining Frequent and Infrequent Features from Chinese Customer Reviews [J]. Journal of Theoretical and Applied Information Technology, 2013, 48(1): 193-199.
[11]JIANG T J, SHIH C W, YANG T H, et al. Enhancement of Feature Engineering for Conditional Random Field Learning in Chinese Word Segmentation Using Unlabeled Data [J]. Computational Linguistics & Chinese Language Processing, 2012,9: 45-86.
[12]祖李军，王卫平. 中文网络评论中提取产品特征的研究[J].计算机系统应用, 2014,23(5):196-201.
ZU Lijun ，WANG Weiping. Research of Extracting Product Features from Chinese Online Reviews [J]. Computer System Applications， 2014,23(5):196-201.
[13]马柏樟，颜志军. 基于潜在狄利特雷分布模型的网络评论产品特征抽取方法[J].计算机集成制造系统, 2014,20(1):96-103.
MA Baizhang，YAN Zhijun. Product Features Extraction of Online Reviews Based on LDA Mode [J]. Computer Integrated Manufacturing Systems，2014,20(1):96-103.
[14]徐建民，王金花，马伟瑜. 利用本体关联度改进的TF-IDF特征词提取方法[J]. 情报科学,2011,29(2):279-283.
XU Jianmin, WANG Jinhua，MA Weiyu. Improved TF-IDF Feature Selection Method Based on Ontology Relative Degree [J]. Intelligence Sciences， 2011,29(2):279-283.
[15]LAFFERTY J, MCCALLUM A, PEREIRA F. Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data[C]// Proceedings of the 18th International Conference on Machine Learning. San Francisco, 2001:282-289.
[16]VELARDI P, STILO G, TOZZI A E, et al. Twitter Mining for Fine-grained Syndromic Surveillance [J]. Artificial Intelligence in Medicine, 2014, 61(3):153-163.
[17]WU Mingxing, WANG Liya, LI Ming, et al. An Approach of Product Usability Evaluation Based on Web Mining in Feature Fatigue Analysis [J]. Computers & Industrial Engineering, 2014,75（1）: 230-238.