Applying HOSVD to Alleviate the Sparsity Problem in Context-aware Recommender Systems

It is quite a great challenge for Context- aware recommender systems (CARS) to generate accurate recommendations with only a few available none-zero con- textual user preferences. This paper presents a new ap- proach to alleviate this high sparsity problem by applying the Higher order singular value decomposition (HOSVD) technique. Firstly, it constructs an N-order tensor to rep- resent multidimensional contextual user preferences and decompose it into (N − 2) 3-order tensors according to different types of context (such as time, location and ac- tivity). Secondly, it introduces HOSVD to automatically discover the latent associations among these different 3- dimensional objects and predicts unknown unidimensional contextual user preferences. Finally, it calculates every contextual influence coefficient that each type of context factor influences user preferences and then constructs a new N-order tensor using weighted linearization method in order to provide recommendations. Experimental eval- uation on a simulated personalized mobile services envi- ronment demonstrates the efficacy of our approach against the other baseline ones.

[1]  Huan Liu,et al.  CubeSVD: a novel approach to personalized Web search , 2005, WWW '05.

[2]  Joos Vandewalle,et al.  A Multilinear Singular Value Decomposition , 2000, SIAM J. Matrix Anal. Appl..

[3]  Mukkai S. Krishnamoorthy,et al.  A random walk method for alleviating the sparsity problem in collaborative filtering , 2008, RecSys '08.

[4]  Xing Xie,et al.  Collaborative Filtering Meets Mobile Recommendation: A User-Centered Approach , 2010, AAAI.

[5]  Aleksandra Mojsilovic,et al.  A Family of Non-negative Matrix Factorizations for One-Class Collaborative Filtering Problems , 2009 .

[6]  Yehuda Koren,et al.  Factor in the neighbors: Scalable and accurate collaborative filtering , 2010, TKDD.

[7]  Panagiotis Symeonidis,et al.  A Unified Framework for Providing Recommendations in Social Tagging Systems Based on Ternary Semantic Analysis , 2010, IEEE Transactions on Knowledge and Data Engineering.

[8]  Bo,et al.  A Novel Collaborative Filtering Using Kernel Methods for Recommender Systems , 2012 .

[9]  Zhang Yujie.  A Cognitive Psychology-Based Approach to User Preferences Elicitation for Mobile Network Services , 2011 .

[10]  Gediminas Adomavicius,et al.  Incorporating contextual information in recommender systems using a multidimensional approach , 2005, TOIS.

[11]  Hsinchun Chen,et al.  Applying associative retrieval techniques to alleviate the sparsity problem in collaborative filtering , 2004, TOIS.

[12]  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.

[13]  Wang Li.  Context-Aware Recommender Systems , 2012 .

[14]  Lu,et al.  Using Non-Negative Matrix Factorization to Cluster Learners and Construct Learning Communities , 2011 .

[15]  Philip S. Yu,et al.  Music Recommendation Using Content and Context Information Mining , 2010, IEEE Intelligent Systems.

[16]  Michael J. Pazzani,et al.  Learning Collaborative Information Filters , 1998, ICML.