A Novel Recommendation Algorithm Based on Heterogeneous Information Network Similarity and Preference Diffusion

Recommender system has been proposed as a key tool to overcome the problem of information overload. In the present era of big data, how to utilization the side information of users, items is a new challenge. This paper put forward a novel solution based on the heterogeneous information network and preference diffusion. The similarity matrices of users and items are initially computed based on meta-path similarity algorithm; three new preference diffusion methods has been proposed to fuse the similarity matrix and the user-item rating matrix; finally uses the traditional recommendation techniques based on matrix factorization to predict the results. With the experiment in a classical data set MovieLens 100 K and the movie attributes extended from IMDb, verifies the effectiveness of the solution that with heterogeneous information network to make full use of users and item attributes information and the preference diffusion with rating matrix can improve the recommendation accuracy effectively.

[1]  Gerhard Friedrich,et al.  Recommender Systems - An Introduction , 2010 .

[2]  Martha Larson,et al.  Collaborative Filtering beyond the User-Item Matrix , 2014, ACM Comput. Surv..

[3]  Joseph S. Valacich,et al.  The Influence of Task Interruption on Individual Decision Making: An Information Overload Perspective , 1999 .

[4]  Nada Lavrac,et al.  A Methodology for Mining Document-Enriched Heterogeneous Information Networks , 2011, Discovery Science.

[5]  John Riedl,et al.  Application of Dimensionality Reduction in Recommender System - A Case Study , 2000 .

[6]  Laks V. S. Lakshmanan,et al.  HeteroMF: recommendation in heterogeneous information networks using context dependent factor models , 2013, WWW.

[7]  Philip S. Yu,et al.  PathSim , 2011, Proc. VLDB Endow..

[8]  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.

[9]  Yizhou Sun,et al.  Full-text based context-rich heterogeneous network mining approach for citation recommendation , 2014, IEEE/ACM Joint Conference on Digital Libraries.

[10]  Yizhou Sun,et al.  Personalized entity recommendation: a heterogeneous information network approach , 2014, WSDM.

[11]  Yizhou Sun,et al.  Mining Heterogeneous Information Networks: Principles and Methodologies , 2012, Mining Heterogeneous Information Networks: Principles and Methodologies.

[12]  Douglas B. Terry,et al.  Using collaborative filtering to weave an information tapestry , 1992, CACM.

[13]  Jianyong Wang,et al.  Incorporating heterogeneous information for personalized tag recommendation in social tagging systems , 2012, KDD.

[14]  Jiawei Han,et al.  Community Mining from Multi-relational Networks , 2005, PKDD.

[15]  Sophie Ahrens,et al.  Recommender Systems , 2012 .

[16]  Philip S. Yu,et al.  HeteSim: A General Framework for Relevance Measure in Heterogeneous Networks , 2013, IEEE Transactions on Knowledge and Data Engineering.

[17]  Jiawei Han.  Mining Heterogeneous Information Networks by Exploring the Power of Links , 2009, Discovery Science.

[18]  Yizhou Sun,et al.  Recommendation in heterogeneous information networks with implicit user feedback , 2013, RecSys.

[19]  Rajeev Motwani,et al.  The PageRank Citation Ranking : Bringing Order to the Web , 1999, WWW 1999.

[20]  H. Sebastian Seung,et al.  Learning the parts of objects by non-negative matrix factorization , 1999, Nature.

[21]  Philip S. Yu,et al.  HeteRecom: a semantic-based recommendation system in heterogeneous networks , 2012, KDD.

[22]  Jennifer Widom,et al.  SimRank: a measure of structural-context similarity , 2002, KDD.

[23]  Philip S. Yu,et al.  Relevance search in heterogeneous networks , 2012, EDBT '12.