A Novel Evidence-Based Bayesian Similarity Measure for Recommender Systems
User-based collaborative filtering, a widely used nearest neighbour-based recommendation technique, predicts an item’s rating by aggregating its ratings from similar users. User similarity is traditionally calculated by cosine similarity or the Pearson correlation coefficient. However, both of these...
Saved in:
Main Authors: | , , |
---|---|
Other Authors: | |
Format: | Article |
Language: | English |
Published: |
2017
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/84847 http://hdl.handle.net/10220/41981 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
id |
sg-ntu-dr.10356-84847 |
---|---|
record_format |
dspace |
spelling |
sg-ntu-dr.10356-848472020-03-07T11:48:57Z A Novel Evidence-Based Bayesian Similarity Measure for Recommender Systems Guo, Guibing Zhang, Jie Yorke-Smith, Neil School of Computer Science and Engineering Bayesian similarity Recommender systems User-based collaborative filtering, a widely used nearest neighbour-based recommendation technique, predicts an item’s rating by aggregating its ratings from similar users. User similarity is traditionally calculated by cosine similarity or the Pearson correlation coefficient. However, both of these measures consider only the direction of rating vectors, and suffer from a range of drawbacks. To overcome these issues, we propose a novel Bayesian similarity measure based on the Dirichlet distribution, taking into consideration both the direction and length of rating vectors. We posit that not all the rating pairs should be equally counted in order to accurately model user correlation. Three different evidence factors are designed to compute the weights of rating pairs. Further, our principled method reduces correlation due to chance and potential system bias. Experimental results on six real-world datasets show that our method achieves superior accuracy in comparison with counterparts. MOE (Min. of Education, S’pore) Accepted version 2017-01-06T05:15:38Z 2019-12-06T15:52:14Z 2017-01-06T05:15:38Z 2019-12-06T15:52:14Z 2016 Journal Article Guo, G., Zhang, J., & Yorke-Smith, N. (2016). A Novel Evidence-Based Bayesian Similarity Measure for Recommender Systems. ACM Transactions on the Web, 10(2), 8-. 1559-1131 https://hdl.handle.net/10356/84847 http://hdl.handle.net/10220/41981 10.1145/2856037 en ACM Transactions on the Web © 2016 Association for Computing Machinery (ACM). This is the author created version of a work that has been peer reviewed and accepted for publication by ACM Transactions on the Web, Association for Computing Machinery (ACM). It incorporates referee’s comments but changes resulting from the publishing process, such as copyediting, structural formatting, may not be reflected in this document. The published version is available at: [http://dx.doi.org/10.1145/2856037]. 30 p. application/pdf |
institution |
Nanyang Technological University |
building |
NTU Library |
country |
Singapore |
collection |
DR-NTU |
language |
English |
topic |
Bayesian similarity Recommender systems |
spellingShingle |
Bayesian similarity Recommender systems Guo, Guibing Zhang, Jie Yorke-Smith, Neil A Novel Evidence-Based Bayesian Similarity Measure for Recommender Systems |
description |
User-based collaborative filtering, a widely used nearest neighbour-based recommendation technique, predicts an item’s rating by aggregating its ratings from similar users. User similarity is traditionally calculated by cosine similarity or the Pearson correlation coefficient. However, both of these measures consider only the direction of rating vectors, and suffer from a range of drawbacks. To overcome these issues, we propose a novel Bayesian similarity measure based on the Dirichlet distribution, taking into consideration both the direction and length of rating vectors. We posit that not all the rating pairs should be equally counted in order to accurately model user correlation. Three different evidence factors are designed to compute the weights of rating pairs. Further, our principled method reduces correlation due to chance and potential system bias. Experimental results on six real-world datasets show that our method achieves superior accuracy in comparison with counterparts. |
author2 |
School of Computer Science and Engineering |
author_facet |
School of Computer Science and Engineering Guo, Guibing Zhang, Jie Yorke-Smith, Neil |
format |
Article |
author |
Guo, Guibing Zhang, Jie Yorke-Smith, Neil |
author_sort |
Guo, Guibing |
title |
A Novel Evidence-Based Bayesian Similarity Measure for Recommender Systems |
title_short |
A Novel Evidence-Based Bayesian Similarity Measure for Recommender Systems |
title_full |
A Novel Evidence-Based Bayesian Similarity Measure for Recommender Systems |
title_fullStr |
A Novel Evidence-Based Bayesian Similarity Measure for Recommender Systems |
title_full_unstemmed |
A Novel Evidence-Based Bayesian Similarity Measure for Recommender Systems |
title_sort |
novel evidence-based bayesian similarity measure for recommender systems |
publishDate |
2017 |
url |
https://hdl.handle.net/10356/84847 http://hdl.handle.net/10220/41981 |
_version_ |
1681041701413584896 |