计算机科学 ›› 2017, Vol. 44 ›› Issue (9): 243-249.doi: 10.11896/j.issn.1002-137X.2017.09.046

• 人工智能 • 上一篇    下一篇

融合时间因素和用户评分特性的协同过滤算法

曾安,高成思,徐小强   

  1. 广东工业大学计算机学院 广州510006,广东工业大学计算机学院 广州510006,广东工业大学计算机学院 广州510006
  • 出版日期:2018-11-13 发布日期:2018-11-13
  • 基金资助:
    本文受国家自然科学基金项目(61300107),广东省自然科学基金项目(S2012010010212),广州市科技计划项目(201504301341059),广东省科技计划项目(2014B090901053)资助

Collaborative Filtering Algorithm Incorporating Time Factor and User Preference Properties

ZENG An, GAO Cheng-si and XU Xiao-qiang   

  • Online:2018-11-13 Published:2018-11-13

摘要: 针对传统协同过滤技术在现实应用中遇到的数据稀疏性问题和局限性,充分挖掘用户评分特性,提出融合时间因素和用户评分特性的协同过滤算法(CF-TP)。引入用户偏好模型,将用户-项目评分矩阵转化为用户-项目偏好得分矩阵,以降低用户评分习惯差异带来的影响。在预测用户对项目的偏好得分时,充分考虑用户之间的非对称影响度,根据用户兴趣随时间的变化引入时间权重函数,以提高top-N推荐的准确率。基于HetRec2011和MovieLens1M数据集的实验结果表明,相对于目前比较流行的算法,所提算法在推荐结果的准确率、召回率、F1值上均有较大的提升,有效提高了推荐系统的推荐质量。

关键词: 推荐系统,协同过滤,时间因素,用户偏好,非对称性

Abstract: To alleviate the impact of data sparsity and overcome the limits of traditional collaborative filtering algorithm,a novel collaborative filtering algorithm(CF-TP) by incorporating time factor and user preference properties was proposed.A user-item rating matrix is firstly converted to a user-item preference matrix by introducing a preference model,and this helps to reduce the impact of different users’ rating habits.Then the asymmetric impact between users is taken into consideration when computing the predicted scores.What’s more important,to further improve the accuracy of top-N recommendation,the time weight function considering uses’ dynamic interest changed with time is designed.The experiment results on HetRec2011 and MovieLens1M data set suggest that the proposed algorithm is superior to other advanced approaches in precision,recall and F1.

Key words: Recommender system,Collaborative filtering,Time factor,User preference,Asymmetric

[1] KHUSRO S,ALI Z,ULLAH I.Recommender Systems:Issues,Challenges,and Research Opportunities[M]∥Information Scien-ce and Applications (ICISA) 2016.Springer Singapore,2016.
[2] KARATZOGLOU A,HIDASI,Bal,et al.RecSys’16 Workshop on Deep Learning for Recommender Systems (DLRS)[C]∥ACM Conference on Recommender Systems.ACM,2016:415-416.
[3] MENG X W,LIU S D,ZHANG Y J,et al.Research on Social Recommender Systems[J].Journal of Software,2015,26(6):1356-1372.(in Chinese) 孟祥武,刘树栋,张玉洁,等.社会化推荐系统研究[J].软件学报,2015,26(6):1356-1372.
[4] VAIRACHILAI S,KAVITHADEVI M K,RAJA M.Allevia-ting the Cold Start Problem in Recommender Systems Based on Modularity Maximization Community Detection Algorithm[J].Circuits and Systems,2016,7(8):1268-1279.
[5] HUANG C G,YIN J,WANG J,et al.Uncertain Neighbors’Collaborative Filtering Recommendation Algorithm[J].Chinese Journal of Computers,2010,33(8):1369-1377.(in Chinese) 黄创光,印鉴,汪静,等.不确定近邻的协同过滤推荐算法[J].计算机学报,2010,33(8):1369-1377.
[6] PIRASTEH P,HWANG D,JUNG J E.Weighted SimilaritySchemes for High Scalability in User-Based Collaborative Filtering[J].Mobile Networks & Applications,2015,20(4):497-507.
[7] LEE J,LEE D,LEE Y C,et al.Improving the accuracy of top- N,recommendation using a preference model[J].Information Sciences,2016,348(c):290-304.
[8] ADOMAVICIUS G,TUZHILIN A.Context-Aware Recommender Systems[C]∥Recommender Systems HandbookSpringer US.2011:217-253.
[9] KARAHODZA B,DONKO D.Feature enhanced time-awarere-commender system[C]∥Xxv International Conference on Information,Communication and Automation Technologies.2015:1-6.
[10] XIA C,JIANG X,LIU S,et al.Dynamic item-based recommendation algorithm with time decay[C]∥International Confe-rence on Natural Computation(Icnc 2010).Yantai,Shandong,China,2010:242-247.
[11] KARAHODZA B,SUPIC H,DONKO D.An approach to design of time-aware recommender system based on changes in group user’s preferences[C]∥X International Symposium on Telecommunications.2014:1-4.
[12] ZHANG Y C,CHEN C.A Collaborative Filtering Algorithm Based on Time Period Partition[C]∥Third International Symposium on Intelligent Information Technology and Security Informatics.IEEE Computer Society,2010:777-780.
[13] CREMONESI P,KOREN Y,TURRIN R.Performance ofrecommender algorithms on top-n recommendation tasks[C]∥ACM Conference on Recommender Systems(Recsys 2010).Barcelona,Spain,2010:39-46.
[14] JIN R,SI L,ZHAI C X,et al.Collaborative filtering with decoupled models for preferences and ratings[C]∥ACM CIKM International Conference on Information and Knowledge Management.New Orleans,Louisiana,Usa,2003:309-316.
[15] ADIBI P,LADANI B T.A collaborative filtering recommender system based on user’s time pattern activity[C]∥Information and Knowledge Technology.2013:252-257.
[16] BOBADILLA J,ORTEGA F,H ERNANDO A,et al.A collaborative filtering approach to mitigate the new user cold start problem[J].Knowledge-Based Systems,2012,26:225-238.

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!