基于在线评论情感分析的改进协同过滤推荐模型

doi:10.6040/j.issn.1672-3961.0.2017.485

山东大学学报 (工学版) ›› 2019, Vol. 49 ›› Issue (1): 47-54.doi: 10.6040/j.issn.1672-3961.0.2017.485

基于在线评论情感分析的改进协同过滤推荐模型

钱春琳^1,²(),张兴芳^3,*(),孙丽华²

1. 河海大学企业管理学院, 江苏常州 213022
2. 天津大学管理与经济学部, 天津 300072
3. 聊城大学数学科学学院, 山东聊城 252000

收稿日期:2017-10-08 出版日期:2019-02-20 发布日期:2019-03-01
通讯作者: 张兴芳 E-mail:qiancl1997@126.com;sunlh68@tju.edu.cn
作者简介:钱春琳(1997—),女,江苏南通人，硕士研究生，主要研究方向为个性化推荐算法研究与在线评论挖掘. E-mail:qiancl1997@126.com
基金资助:
国家自然科学基金(11471152)

Advanced collaborative filtering recommendation model based on sentiment analysis of online review

Chunlin QIAN^1,²(),Xingfang ZHANG^3,*(),Lihua SUN²

1. School of Business Administration, Hohai University, Changzhou 213022, Jiangsu, China
2. College of Management and Economics, Tianjin University, Tianjin 300072, China
3. School of Mathematical Sciences, Liaocheng University, Liaocheng 252000, Shandong, China

Received:2017-10-08 Online:2019-02-20 Published:2019-03-01
Contact: Xingfang ZHANG E-mail:qiancl1997@126.com;sunlh68@tju.edu.cn
Supported by:
国家自然科学基金(11471152)

摘要/Abstract

摘要：

针对在线中文评论中用户主观意见的不确定性,提出一种基于不确定理论的情感分析模型,并结合情感分析模型设计了个性化推荐算法。采用分词工具ICTCLAS和IKAnalyzer预处理在线中文评论,并基于情感词典(HowNet)计算特征词的点互信息值;应用不确定变量与不确定集设计情感分析模型;根据情感分析模型设计新的最近邻居搜索方法并产生推荐。在两个真实数据集上进行试验,试验结果表明,该方法能够有效改进推荐结果的准确率,缓解数据稀疏问题。

关键词: 推荐模型, 不确定变量, 不确定集, 在线评论, 情感分析

Abstract:

Aiming at the uncertainty of users' subject opinions in online Chinese review, a sentiment analysis model was proposed based on uncertainty theory. An individual recommendation algorithm was designed on the basis of the proposed sentiment analysis model. Firstly, the tokenizers of ICTCLAS and IKAnalyzer were used to preprocess online Chinese review to generate characteristic words, and the point mutual information value of characteristic words accounting for the sentiment direction were computed based on sentiment dictionary (HowNet). Then, the sentiment analysis model was established via uncertainty theory of uncertain variable and uncertain set. In addition, the new similarity formula based on the proposed model was used to search the nearest neighbors. Finally, the recommendation lists were given. The experiments were carried out on two real datasets. The results showed that the proposed method could effectively improve the accuracy of recommendation and alleviate the sparse data problem.

Key words: recommendation model, uncertain variable, uncertain set, online review, sentiment analysis

中图分类号:

TP391

钱春琳,张兴芳,孙丽华. 基于在线评论情感分析的改进协同过滤推荐模型[J]. 山东大学学报 (工学版), 2019, 49(1): 47-54.

Chunlin QIAN,Xingfang ZHANG,Lihua SUN. Advanced collaborative filtering recommendation model based on sentiment analysis of online review[J]. Journal of Shandong University(Engineering Science), 2019, 49(1): 47-54.

图/表 8

图1

表1

表2

表3

表4

表5

图2

图3

参考文献 17

1	RESNICK P , VARIAN H R . Recommender systems[J]. Commun ACM, 1997, 40 (3): 56- 58.
2	DE CAMPOS L M , FERNÁNDEZ-LUNA J M , HUETE J F . A collaborative recommender system based on probabilistic inference from fuzzy observations[J]. Fuzzy Sets and Systems, 2008, 159 (12): 1554- 1576. doi: 10.1016/j.fss.2008.01.016
3	MURTHI B P S , SARKAR S . The Role of the management sciences in research on personalization[J]. Management Science, 2003, 49 (10): 1344- 1362. doi: 10.1287/mnsc.49.10.1344.17313
4	WANG J H, LIU T W. Improving sentiment rating of movie review comments for recommendation[C]//Consumer Electronics-Taiwan (ICCE-TW), 2017 IEEE International Conference on. Taipei, China: IEEE, 2017: 433-434.
5	CHEN L , CHEN G , WANG F . Recommender systems based on user reviews: the state of the art[J]. User Modeling and User-Adapted Interaction, 2015, 25 (2): 99- 154. doi: 10.1007/s11257-015-9155-5
6	QIU L , GAO S , CHENG W , et al. Aspect-based latent factor model by integrating ratings and reviews for recommender system[J]. Knowledge-Based Systems, 2016, 110, 233- 243. doi: 10.1016/j.knosys.2016.07.033
7	BAO Y, FANG H, ZHANG J. TopicMF: simultaneously exploiting ratings and reviews for recommendation[C]//Proceedings of the Twenty-Eighth AAAI Conference on Artificial Intelligence. Québec, Canada: AAAI, 2014, 14: 2-8.
8	ZHAI C, PENG J. Mining latent features from reviews and ratings for item recommendation[C]//Computational Science and Computational Intelligence (CSCI), 2016 International Conference on. Las Vegas, USA: IEEE, 2016: 1119-1125.
9	ZHANG Y, LAI G, ZHANG M, et al. Explicit factor models for explainable recommendation based on phrase-level sentiment analysis[C]//Proceedings of the 37th international ACM SIGIR conference on Research & development in information retrieval. Gold Coast, Queensland, Australia: ACM, 2014: 83-92.
10	KOOHI H , KIANI K . User based collaborative filtering using fuzzy C-means[J]. Measurement, 2016, 91, 134- 139. doi: 10.1016/j.measurement.2016.05.058
11	LIU B . Uncertainty theory[M]. Berlin, Germany: Springer, 2007.
12	LIU B . Uncertainty theory-a branch of mathematics for modeling human uncertainty[M]. Berlin, Germany: Springer, 2010.
13	DONG Zhendong , DONG Qiang . Hownet and the computation of meaning (with CD-ROM)[M]. Singapore: World Scientific Publishing Company, 2006.
14	TURNEY P D. Thumbs up or thumbs down? semantic orientation applied to unsupervised classification of reviews[C]//Proceedings of the 40th annual meeting on association for computational linguistics. Philadelphia, USA: Association for Computational Linguistics, 2002: 417-424.
15	LIU B . Uncertain logic for modeling human language[J]. Journal of Uncertain Systems, 2011, 5 (1): 3- 20.
16	WANG X , GAO Z , GUO H . Delphi method for estimating uncertainty distributions[J]. International Journal on Information, 2012, 15 (2): 449- 459.
17	SARWAR B, KARYPIS G, KONSTAN J, et al. Item-based collaborative filtering recommendation algorithms[C]//Proceedings of the 10th international conference on World Wide Web. Hong Kong, China: ACM, 2001: 285-295.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed

等级1	等级2	等级3	等级4	等级5	等级6
不得了	着实	更加	点点滴滴	半点	不为过
极度	太甚	还要	好生	不堪	开外
酷	不胜	愈加	稍稍	轻度	出头
莫大	特别	这样	有点儿	相对	过
入骨	多多	越是	相当	不怎么	过分

负向情感词	正向情感词
油腻、虚假、讨厌、脆弱、陈旧	接受、暖心、认可、新鲜、最爱

	记录数	用户数	项目数	稀疏度/%
餐馆数据集	560 113	9 171	2 903	97.90
酒店数据集	10 372	1 411	528	98.61
手机数据集	219 732	215 402	1 007	99.90

邻居数量	USR-CF		CF
邻居数量	Precision	F₁	Precision	F₁
5	0.020 48	0.021 17	0.015 26	0.010 05
15	0.018 02	0.020 52	0.013 45	0.008 84
25	0.016 88	0.018 82	0.012 31	0.007 87
35	0.014 91	0.016 91	0.010 50	0.006 69
45	0.013 74	0.015 07	0.010 66	0.007 11
55	0.011 06	0.013 59	0.009 68	0.006 21

推荐列表		USR-CF		FC-Means
推荐列表		Precision	F₁	Precision	F₁
	5	0.018 19	0.017 78	0.017 40	0.015 66
	10	0.018 02	0.020 52	0.018 00	0.017 36
餐馆数据集	15	0.020 70	0.023 83	0.019 04	0.018 63
	20	0.021 94	0.024 78	0.020 01	0.019 97
	25	0.022 30	0.025 61	0.020 69	0.021 31
	5	0.004 81	0.004 87	0.003 81	0.003 97
	10	0.004 80	0.005 67	0.003 84	0.004 31
手机数据集	15	0.004 80	0.005 75	0.003 88	0.004 40
	20	0.004 82	0.005 33	0.003 90	0.004 40
	25	0.004 75	0.005 06	0.003 90	0.004 47

基于在线评论情感分析的改进协同过滤推荐模型

Advanced collaborative filtering recommendation model based on sentiment analysis of online review

RichHTML

PDF (PC)

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 17

相关文章 8

多维度评价

本文评价

推荐阅读 10

[1]	杨修远,彭韬,杨亮,林鸿飞. 基于知识蒸馏的自适应多领域情感分析[J]. 山东大学学报 (工学版), 2021, 51(3): 15-21.
[2]	蔡国永,贺歆灏,储阳阳. 基于空间注意力和卷积神经网络的视觉情感分析[J]. 山东大学学报 (工学版), 2020, 50(4): 8-13.
[3]	蔡国永, 林强, 任凯琪. 基于域对抗网络和BERT的跨领域文本情感分析[J]. 山东大学学报 (工学版), 2020, 50(1): 1-7.
[4]	周荣翔,贾修一. 中文反语识别特征分析[J]. 山东大学学报 (工学版), 2019, 49(1): 41-46.
[5]	沈冀,马志强,李图雅,张力. 面向短文本情感分析的词扩充LDA模型[J]. 山东大学学报(工学版), 2018, 48(3): 120-126.
[6]	周哲, 商琳. 一种基于动态词典和三支决策的情感分析方法[J]. 山东大学学报(工学版), 2015, 45(1): 19-23.
[7]	周咏梅1,阳爱民1,林江豪2. 中文微博情感词典构建方法[J]. 山东大学学报(工学版), 2014, 44(3): 36-40.
[8]	周咏梅1，杨佳能2,阳爱民2. 面向文本情感分析的中文情感词典构建方法[J]. 山东大学学报(工学版), 2013, 43(6): 27-33.