Challenging the long tail recommendation

Yin, Hongzhi, Cui, Bin, Li, Jing, Yao, Junjie and Chen, Chen (2012). Challenging the long tail recommendation. In: Proceedings of the 38th International Conference on Very Large Data Bases. 38th International Conference on Very Large Data Bases 2012, (VLDB 2012), Istanbul, Turkey, (896-907). 27-31 August 2012. doi:10.14778/2311906.2311916

Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads

Author Yin, Hongzhi
Cui, Bin
Li, Jing
Yao, Junjie
Chen, Chen
Title of paper Challenging the long tail recommendation
Conference name 38th International Conference on Very Large Data Bases 2012, (VLDB 2012)
Conference location Istanbul, Turkey
Conference dates 27-31 August 2012
Proceedings title Proceedings of the 38th International Conference on Very Large Data Bases   Check publisher's open access policy
Journal name Proceedings of the VLDB Endowment International Conference on Very Large Data Bases   Check publisher's open access policy
Place of Publication New York, NY United States
Publisher Association for Computing Machinery
Publication Year 2012
Year available 2012
Sub-type Fully published paper
DOI 10.14778/2311906.2311916
Open Access Status Not yet assessed
ISSN 2150-8097
Volume 5
Issue 9
Start page 896
End page 907
Total pages 12
Language eng
Formatted Abstract/Summary
The success of “infinite-inventory” retailers such as Amazon.com and Netflix has been largely attributed to a “long tail” phenomenon.  Although the majority of their inventory is not in high demand, these niche products, unavailable at limited-inventory competitors, generate a significant fraction of total revenue in aggregate. In addition, tail product availability can boost head sales by offering consumers the convenience of “one-stop shopping” for both their mainstream and niche tastes. However, most of existing recommender systems, especially collaborative filter based methods, can not recommend tail products due to the data sparsity issue. It has been widely acknowledged that to recommend popular products is easier yet more trivial while to recommend long tail products adds more novelty yet it is also a more challenging task.   In this paper, we propose a novel suite of graph-based algorithms for the long tail recommendation. We first represent user-item information
with undirected edge-weighted graph and investigate the theoretical foundation of applying Hitting Time algorithm for long
tail item recommendation. To improve recommendation diversity and accuracy, we extend Hitting Time and propose efficient Absorbing Time algorithm to help users find their favorite long tail items.  Finally, we refine the Absorbing Time algorithm and propose two entropy-biased Absorbing Cost algorithms to distinguish the variation on different user-item rating pairs, which further enhances the effectiveness of long tail recommendation.  Empirical  experiments on two real life datasets show that our proposed algorithms are effective to recommend long tail items and outperform state-of-the-art recommendation techniques.
Q-Index Code E1
Q-Index Status Provisional Code
Institutional Status Non-UQ

Document type: Conference Paper
Sub-type: Fully published paper
Collections: School of Information Technology and Electrical Engineering Publications
School of Biomedical Sciences Publications
 
Versions
Version Filter Type
Citation counts: TR Web of Science Citation Count  Cited 34 times in Thomson Reuters Web of Science Article | Citations
Scopus Citation Count Cited 62 times in Scopus Article | Citations
Google Scholar Search Google Scholar
Created: Sun, 01 Mar 2015, 22:05:09 EST by Hongzhi Yin on behalf of School of Information Technol and Elec Engineering