Please wait a minute...
投稿  |   English  | 
 
高级检索
   首页  |  最新收录  |  当期目录  |  过刊浏览  |  作者中心  |  关于期刊   开放获取  
投稿  |   English  | 
Engineering    2016, Vol. 2 Issue (2) : 212-224     https://doi.org/10.1016/J.ENG.2016.02.013
Research |
非独立同分布推荐系统:推荐范式转换的综述和框架
Cao Longbing()
Advanced Analytics Institute, University of Technology Sydney, Sydney, NSW 2007, Australia
全文: PDF(2275 KB)   HTML
导出: BibTeX | EndNote | Reference Manager | ProCite | RefWorks    
摘要 

虽然推荐系统在我们的生活、学习、工作和娱乐中扮演着越来越重要的角色,但是很多时候我们收到的推荐都是不相关的、重复的,或者包含不感兴趣的产品和服务。这些差的推荐系统产生的原因来源于一个本征假设:传统的理论和推荐系统认为用户和物品是独立同分布的(IID)。另一个明显的现象是,虽然投入了很多的精力模拟用户或者物品的特殊属性,但用户和物品的总体属性及它们之间的非独立同分布性(non-IID) 被忽略了。本文先讨论了推荐系统的非独立同分布性,紧接着介绍了非独立同分布性原理,目的是从耦合和异构性的角度来深入阐述传统的推荐系统的固有本质。这种非独立同分布推荐系统引起了传统推荐系统范式的转化—— 从独立同分布向非独立同分布进行转化,希望能够形成高效的、相关性高的、个人订制和可操作的推荐系统。这种系统创造了令人兴奋的能够解决包含冷启动、以稀疏数据为基础、跨域、基于群组信息和欺诈攻击等各种复杂情况的新的研究方向和解决方案。

Abstract

While recommendation plays an increasingly critical role in our living, study, work, and entertainment, the recommendations we receive are often for irrelevant, duplicate, or uninteresting products and services. A critical reason for such bad recommendations lies in the intrinsic assumption that recommended users and items are independent and identically distributed (IID) in existing theories and systems. Another phenomenon is that, while tremendous efforts have been made to model specific aspects of users or items, the overall user and item characteristics and their non-IIDness have been overlooked. In this paper, the non-IID nature and characteristics of recommendation are discussed, followed by the non-IID theoretical framework in order to build a deep and comprehensive understanding of the intrinsic nature of recommendation problems, from the perspective of both couplings and heterogeneity. This non-IID recommendation research triggers the paradigm shift from IID to non-IID recommendation research and can hopefully deliver informed, relevant, personalized, and actionable recommendations. It creates exciting new directions and fundamental solutions to address various complexities including cold-start, sparse data-based, cross-domain, group-based, and shilling attack-related issues.

Keywords Independent and identically distributed (IID)      Non-IID      Heterogeneity      Coupling relationship      Coupling learning      Relational learning      IIDness learning      Non-IIDness learning      Recommender system      Recommendation      Non-IID recommendation     
通讯作者: Cao Longbing     E-mail: longbing.cao@gmail.com
最新录用日期:    发布日期: 2016-07-04
服务
推荐给朋友
免费邮件订阅
RSS订阅
作者相关文章
Longbing Cao
引用本文:   
Longbing Cao. Non-IID Recommender Systems: A Review and Framework of Recommendation Paradigm Shifting[J]. Engineering, 2016, 2(2): 212-224.
网址:  
http://journal.hep.com.cn/eng/EN/10.1016/J.ENG.2016.02.013     OR     http://journal.hep.com.cn/eng/EN/Y2016/V2/I2/212
Fig.1  A systematic view of recommendation.
Fig.2  Non-IIDness in recommendation.
Fig.3  The multilayer model of recommendation research.
Fig.4  The four generations of recommendation research.
Fig.5  Framework for non-IID recommendation.
Data setMetricsUBCF(Improve)IBCF(Improve)CMF
MovielensMAE0.9027 (0.49%)0.9220 (2.42%)0.8978
RMSE1.0022 (0.18%)1.1958 (19.54%)1.0004
BookcrossingMAE1.8064 (33.02%)1.7865 (31.03%)1.4762
RMSE3.9847 (24.68%)3.9283 (19.04%)3.7379
Tab.1  CMF versus CF on Movielens and Bookcrossing.
1 JannachD, ZankerM, FelfernigA, FriedrichG. Recommender systems: an introduction. Cambridge: Cambridge University Press; 2010.
2 RicciF, RokachL, ShapiraB, KantorPB, editors. Recommender systems handbook. 2nd ed. New York: Springer; 2015.
3 CaoL. Data science: a comprehensive overview. Technical report. Sydney: University of Technology Sydney; 2016.
4 McKinsey Global Institute;ManyikaJ, ChuiM, BrownB, BughinJ, DobbsR, RoxburghC, Big data: the next frontier for innovation, competition, and productivity. New York: McKinsey Global Institute; 2011.
5 CaoL. Non-IIDness learning in behavioral and social data. Comput J2014;57(9):1358–70.
6 CaoL. Coupling learning of complex interactions. Inform Process Manag2015;51(2):167–86.
7 CaoL. In-depth behavior understanding and use: the behavior informatics approach. Inform Sciences2010;180(17):3067–85.
8 CaoL. YuPS, editors. Behavior computing: modeling, analysis, mining and decision. London: Springer; 2012.
9 FuB, XuG, CaoL, WangZ, WuZ. Coupling multiple views of relations for recommendation. In: Cao T, Lim EP, Zhou ZH, Ho TB, Cheung D, Motoda H, editors Advances in Knowledge Discovery and Data mining: 19th Pacific-Asia Conference, Part II; 2015 May 19-22; Ho Chi Minh City, Vietnam. Switzerland: Springer International Publishing; 2015. p. 723–43.
10 LiT, LuJ,?LópezLM. Preface: intelligent techniques for data science. Int J Intell Syst2015;30(8):851–3.
11 YuY, WangC, GaoY, CaoL, ChenX. A coupled clustering approach for items recommendation. In: Pei J, Tseng VS, Cao L, Motoda H, Xu G, editors Advances in Knowledge Discovery and Data mining: 17th Pacific-Asia Conference, Part II; 2013 Apr 14-17; Gold Coast, Australia. Heidelberg: Springer; 2013. p. 365–76.
12 CaoL, YuPS. Non-IID recommendation theories and systems. IEEE Intell Syst2016;31(2):81–4.
13 CaoL. Data science and analytics: a new era. Int J Data Sci Analyt2016;1(1):1–2.
14 CaoL. Data science: intrinsic challenges and directions. Technical report. Sydney: University of Technology Sydney; 2016.
15 CaoL. Data science: nature and pitfalls. Technical report. Sydney: University of Technology Sydney; 2016.
16 SuX, KhoshgoftaarTM. A survey of collaborative filtering techniques. Adv Artif Intell2009;2009(4):1–19.
17 KorenY. Factorization meets the neighborhood: a multifaceted collaborative filtering model. In: Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; 2008 Aug 24-27; New York, USA;?2008. p. 426–34.
18 SarwarB, KarypisG, KonstanJ, RiedlJ. Item-based collaborative filtering recommendation algorithms. In: Proceedings of the 10th International Conference on the World Wide Web; 2001 May 1-5; Hong Kong, China; 2001. p. 285–95.
19 DeshpandeM, KarypisG. Item-based top-N recommendation algorithms. ACM Trans Inform Syst2004;22(1):143–77.
20 MaH, YangH, LyuMR, KingI. SoRec: social recommendation using probabilistic matrix factorization. In: Proceedings of the 17th ACM Conference on Information and Knowledge Management; 2008 Oct 26-30; Napa Valley, CA, USA; 2008. p. 931–40.
21 MaH. An experimental study on implicit social recommendation. In: Proceedings of the 36th International ACM SIGIR conference on Research and Development in Information Retrieval; 2013 Jul 28-Aug 1; Dublin, Ireland; 2013. p. 73–82.
22 SinghAP, GordonGJ. Relational learning via collective matrix factorization. In: Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; 2008 Aug 24-27; Las Vegas, NV, USA; 2008. p. 650–8.
23 HuL, CaoJ, XuG, CaoL, GuZ, CaoW. Deep modeling of group preferences for group-based recommendation. In: Proceedings?of?the 28th AAAI Conference on Artificial Intelligence; 2014 Jul 27-31;?Québec City, Canada; 2014. p. 1861–7.
24 HuL, CaoJ, XuG, WangJ, GuZ, CaoL. Cross-domain collaborative filtering via bilinear multilevel analysis. In: Proceedings of the 23rd International Joint Conference on Artificial Intelligence; 2013 Aug 3-9; Beijing, China; 2013. p. 1–7.
25 HuL, CaoJ, XuG, CaoL, GuZ, ZhuC. Personalized recommendation via cross-domain triadic factorization. In:?Proceedings of the 22nd International Conference on World Wide Web; 2013 May 13-17; Rio de Janeiro, Brazil; 2013. p. 595–606.
26 YangX, SteckH, LiuY. Circle-based recommendation in online social networks. In: Proceedings of the 18th ACM SIGKDD Knowledge Discovery and Data Mining; 2012 Aug 12-16; Beijing, China; 2012. p. 1267–75.
27 LiF, XuG, CaoL, FanX, NiuZ. CGMF: coupled group-based matrix factorization for recommender system. In: Lin X, Manolopoulos Y, Srivastava D, Huang G, editors Web Information Systems Engineering—WISE 2013: 14th International Conference, Part I; 2013 Oct 13-15; Nanjing, China. Heidelberg: Springer. 2013. p. 189–98.
28 BreeseJS, HeckermanD, KadieC. Empirical analysis of predictive algorithms for collaborative filtering. In:?Proceedings of the 14th Conference on Uncertainty in Artificial Intelligence; 1998 Jul 24-26; Madison, WI, USA. San Francisco: Morgan Kaufmann Publishers Inc.; 1998. p. 43–52.
29 ResnickP, IacovouN, SuchakM, BergstromP, RiedlJ. GroupLens: an open architecture for collaborative filtering of netnews. In: Proceedings of ACM 1994 Conference on Computer Supported Cooperative Work; 1994 Oct 22-26; Chapel Hill, NC, USA; 1994. p. 175–86.
30 AlterO, BrownPO, BotsteinD. Singular value decomposition for genome-wide expression data processing and modeling. Proc Natl Acad Sci USA2000;97(18):10101–6.
31 SalakhutdinovR, MnihA. Probabilistic matrix factorization. In:?Platt?JC,?Koller D,?Singer?Y,?Roweis ST, editors Proceedings of ?the 21st Annual Conference on Neural Information Processing Systems 2007; 2007 Dec 3-6; Vancouver, Canada; 2007. p. 1257–64.
32 BurkeR. Hybrid web recommender systems. In:?Brusilovsky P, Kobsa A, Nejdl W, editors The adaptive web. Heidelberg: Springer; 2007. p. 377–408.
33 BurkeR. Hybrid recommender systems: survey and experiments. User Model User-Adapt Interac2002;12(4):331–70.
34 LvLL,?MedoM,?YeungCH,?ZhangYC,?ZhangZK,?ZhouT. Recommender systems. Phys Rep2012;519(1):1–49.
35 KonstanJA, RiedlJ. Recommender systems: from algorithms to user experience. User Model User-Adapt Interact2012;22(1):101–23.
36 BobadillaJ, OrtegaF, HernandoA, GutiérrezA. Recommender systems survey. Knowl-Based Syst2013;46:109–32.
37 ParkDH, KimHK, ChoiIY, KimJK. A literature review and classification of recommender systems research. Expert Syst Appl2012;39(11):10059–72.
38 CaoL, OuY, YuPS. Coupled behavior analysis with applications. IEEE Trans Knowl Data Eng2012;24(8):1378–92.
39 WangC, DongX, ZhouF, CaoL, ChiCH. Coupled attribute similarity learning on categorical data. IEEE Trans Neural Netw Learn Syst2015;26(4):781–97.
40 WangC, CaoL, WangM, LiJ, WeiW, OuY. Coupled nominal similarity in unsupervised learning. In: Proceedings of the 20th ACM Conference on Information and Knowledge Management; 2011 Oct 24-28; Glasgow, UK; 2011. p. 973–8.
41 ChenL, ZengW, YuanQ. A unified framework for recommending items, groups and friends in social media environment via mutual resource fusion. Expert Syst Appl2013;40(8):2889–903.
42 NadeeW. Modeling user profiles for recommender systems [dissertation]. Brisbane: Queensland University of Technology; 2016.
43 LiR, WangS, DengH, WangR, ChangKCC. Towards social user profiling: unified and discriminative influence model for inferring home locations. In: Proceedings of the 18th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; 2012 Aug 12-16; Beijing, China; 2012. p. 1023–31.
44 PopesculA, UngarLH, PennockDM, LawrenceS. Probabilistic models for unified collaborative and content-based recommendation in sparse-data environments. In: Proceedings of the 17th Conference in Uncertainty in Artificial Intelligence; 2001 Aug 2-5; Seattle, WA, USA. San Francisco: Morgan Kaufmann Publishers Inc.; 2001. p. 437–44.
45 ChenQ, HuL, XuJ, LiuW, CaoL. Document similarity analysis via involving both explicit and implicit semantic couplings. In: Proceedings of IEEE Data Science and Advanced Analytics 2015; 2015 Oct 19-21; Paris, France; 2015. p. 1–10.
46 JiangM, CuiP, ChenX, WangF, ZhuW, YangS. Social recommendation with cross-domain transferable knowledge. IEEE Trans Knowl Data Eng2015;27(11):3084–97.
47 PanW, YangQ. Transfer learning in heterogeneous collaborative filtering domains. Artif Intell2013;197:39–55.
48 CaoL. Metasynthetic computing and engineering of complex systems. London: Springer-Verlag; 2015.
49 SonLH. Dealing with the new user cold-start problem in recommender systems: a comparative review. Inform Syst2016;58:87–104.
50 GantnerZ, DrumondL, FreudenthalerC, RendleS, Schmidt-ThiemeL. Learning attribute-to-feature mappings for cold-start recommen-dations. In: Proceedings of the 10th IEEE International Conference on Data Mining; 2010 Dec 13-17; Sydney, Australia; 2010. p. 176–85.
51 MirbakhshN, LingCX. Improving top-N recommendation for cold-start users via cross-domain information. ACM Trans Knowl Discov Data2015;9(4):33.
52 LikaB, KolomvatsosK, HadjiefthymiadesS. Facing the cold start problem in recommender systems. Expert Syst Appl2014;41(4):2065–73.
53 GaoH, TangJ, LiuH. Addressing the cold-start problem in location recommendation using geo-social correlations. Data Min Knowl Disc2015;29(2):299–323.
54 GunesI, KaleliC, BilgeA, PolatH. Shilling attacks against recommender systems: a comprehensive survey. Artif Intell Rev2014;42(4):767–99.
55 PangG, CaoL, ChenL. Outlier detection in complex categorical data by modelling the feature value couplings. In: Proceedings of the 25th International Joint Conference on Artificial Intelligence 2016; 2016 Jul 9-15; New York, NY, USA; 2016. p. 1–7.
56 HidasiB, TikkD. General factorization framework for context-aware recommendations. Data Min Knowl Disc2016;30(2):342–71.
57 JackoJA, editor. The human-computer interaction handbook: fundamentals, evolving technologies and emerging applications. <Date>3rd ed</Date>. Boca Raton: CRC Press; 2006.
58 QianXS, YuJY, DaiRW. A new discipline of science—the study of open complex giant system and its methodology. Chin J Syst Eng Electron1993;4(2):2–12.
59 LiuX, NielekR, AdamskaP, WierzbickiA, AbererK. Towards a highly effective and robust Web credibility evaluation system. Decis Support Syst. 2015;79:99–108.
60 AldhahriE, ShandilyaV, ShivaS. Towards an effective crowdsourcing recommendation system: a survey of the state-of-the-art. In: Proceedings of the 2015 IEEE Symposium on Service-Oriented System Engineering; 2015 Mar 30-Apr 3; San Francisco Bay, CA, USA; 2015. p. 372–7.
No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
国内刊号:CN10-1244/N    国际刊号:ISSN2095-8099
版权所有 © 2015 高等教育出版社  《中国工程科学》杂志社
京ICP备11030251号-2

 Engineering