List-Wise Diffusion-Based Recommender Algorithm from Implicit Feedback

Recently, some physical dynamics, including heat conduction and mass diffusion, have found their applications in personalized recommendation. These kinds of nature-inspired approaches have been demonstrated to be both highly efficient and of low computational complexity. However, most of them rely only on the connections between users and objects, but does not take into consideration the sequence of user-object collecting activities. In this paper, the temporal information of users' object-collecting activities is adopted to measure the user-user similarity. we propose a list-wise diffusion-based recommender algorithm, which assigns the user-user similarity as the weight to the links of user-object bipartite network. Experimental results on two benchmark datasets show that our proposed approach can not only enhance the accuracy, but also largely provide more diverse recommendations.

[1]  Gediminas Adomavicius,et al.  Toward the next generation of recommender systems: a survey of the state-of-the-art and possible extensions , 2005, IEEE Transactions on Knowledge and Data Engineering.

[2]  Thomas Hofmann,et al.  Latent semantic models for collaborative filtering , 2004, TOIS.

[3]  Yi-Cheng Zhang,et al.  Bipartite network projection and personal recommendation. , 2007, Physical review. E, Statistical, nonlinear, and soft matter physics.

[4]  Alan Hanjalic,et al.  List-wise learning to rank with matrix factorization for collaborative filtering , 2010, RecSys '10.

[5]  Dominik Endres,et al.  A new metric for probability distributions , 2003, IEEE Transactions on Information Theory.

[6]  Yi-Cheng Zhang,et al.  Effect of initial configuration on network-based recommendation , 2007, 0711.2506.

[7]  Tie-Yan Liu,et al.  Listwise Collaborative Filtering , 2015, SIGIR.

[8]  Qiang Guo,et al.  Information filtering via biased heat conduction , 2011, Physical review. E, Statistical, nonlinear, and soft matter physics.

[9]  Tie-Yan Liu,et al.  Learning to rank: from pairwise approach to listwise approach , 2007, ICML '07.

[10]  Yi Xiong,et al.  List-wise probabilistic matrix factorization for recommendation , 2014, Inf. Sci..

[11]  Yi-Cheng Zhang,et al.  Heat conduction process on community networks as a recommendation model. , 2007, Physical review letters.

[12]  Greg Linden,et al.  Amazon . com Recommendations Item-to-Item Collaborative Filtering , 2001 .

[13]  Yi-Cheng Zhang,et al.  Solving the apparent diversity-accuracy dilemma of recommender systems , 2008, Proceedings of the National Academy of Sciences.

[14]  Jonathan L. Herlocker,et al.  Evaluating collaborative filtering recommender systems , 2004, TOIS.

[15]  J. Marden.  Analyzing and Modeling Rank Data , 1996 .

[16]  Mouzhi Ge,et al.  Beyond accuracy: evaluating recommender systems by coverage and serendipity , 2010, RecSys '10.