Improving the Performance of Recommender System by Exploiting the Categories of Products

In the literature, collaborative filtering (CF) approach and its variations have been proposed for building recommender systems. In CF, recommendations for a given user are computed based on the ratings of k nearest neighbours. The nearest neighbours of target user are identified by computing the similarity between the product ratings of the target user and the product ratings of every other user. In this paper, we have proposed an improved approach to compute the neighborhood by exploiting the categories of products. In the proposed approach, ratings given by a user are divided into different sub-groups based on the categories of products. We consider that the ratings of each sub-group are given by a virtual user. For a target user, the recommendations of the corresponding virtual user are computed by employing CF. Next, the recommendations of the corresponding virtual users of the target user are combined for recommendation. The experimental results on MovieLens dataset show that the proposed approach improves the performance over the existing CF approach.

[1]  Jun Wang,et al.  Unifying user-based and item-based collaborative filtering approaches by similarity fusion , 2006, SIGIR.

[2]  Robert A. Legenstein,et al.  Combining predictions for accurate recommender systems , 2010, KDD.

[3]  John Riedl,et al.  An Empirical Analysis of Design Choices in Neighborhood-Based Collaborative Filtering Algorithms , 2002, Information Retrieval.

[4]  Pearl Pu,et al.  A recursive prediction algorithm for collaborative filtering recommender systems , 2007, RecSys '07.

[5]  Christos Faloutsos,et al.  TANGENT: a novel, 'Surprise me', recommendation algorithm , 2009, KDD.

[6]  Martha Larson,et al.  Exploiting user similarity based on rated-item pools for improved user-based collaborative filtering , 2009, RecSys '09.

[7]  Amnon Meisels,et al.  Ensemble methods for improving the performance of neighborhood-based collaborative filtering , 2009, RecSys '09.

[8]  Li Chen,et al.  A cross-cultural user evaluation of product recommender interfaces , 2008, RecSys '08.

[9]  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.

[10]  Philip S. Yu,et al.  Horting hatches an egg: a new graph-theoretic approach to collaborative filtering , 1999, KDD '99.

[11]  Quan Yuan,et al.  Boosting collaborative filtering based on statistical prediction errors , 2008, RecSys '08.

[12]  John Riedl,et al.  Analysis of recommendation algorithms for e-commerce , 2000, EC '00.

[13]  John Riedl,et al.  Item-based collaborative filtering recommendation algorithms , 2001, WWW '01.

[14]  George Karypis,et al.  Item-based top-N recommendation algorithms , 2004, TOIS.

[15]  P. Krishna Reddy,et al.  Using lower-bound similarity to enhance the performance of recommender systems , 2011, Bangalore Compute Conf..

[16]  Jun Wang,et al.  Unified relevance models for rating prediction in collaborative filtering , 2008, TOIS.

[17]  Cong Yu,et al.  Constructing and exploring composite items , 2010, SIGMOD Conference.

[18]  Jonathan L. Herlocker,et al.  A collaborative filtering algorithm and evaluation metric that accurately model the user experience , 2004, SIGIR '04.

[19]  Gerard Salton,et al.  Research and Development in Information Retrieval , 1982, Lecture Notes in Computer Science.

[20]  Yehuda Koren,et al.  Factorization meets the neighborhood: a multifaceted collaborative filtering model , 2008, KDD.