On top-k recommendation using social networks

Xiwang Yang, Harald Steck, Yang Guo, Yong Liu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Recommendation accuracy can be improved by incorporating trust relationships derived from social networks. Most recent work on social network based recommendation is focused on minimizing the root mean square error (RMSE). Social network based top-k recommendation, which recommends to a user a small number of items at a time, is not well studied. In this paper, we conduct a comprehensive study on improving the accuracy of top-k recommendation using social networks. We first show that the existing social-trust enhanced Matrix Factorization (MF) models can be tailored for top-k recommendation by including observed and missing ratings in their training objective functions. We also propose a Nearest Neighbor (NN) based top-k recommendation method that combines users' neighborhoods in the trust network with their neighborhoods in the latent feature space. Experimental results on two publicly available datasets show that social networks can significantly improve the top-k hit ratio, especially for cold start users. Surprisingly, we also found that the technical approach for combining feedback data (e.g. ratings) with social network information that works best for minimizing RMSE works poorly for maximizing the hit ratio, and vice versa.

Original languageEnglish (US)
Title of host publicationRecSys'12 - Proceedings of the 6th ACM Conference on Recommender Systems
Pages67-74
Number of pages8
DOIs
StatePublished - 2012
Event6th ACM Conference on Recommender Systems, RecSys 2012 - Dublin, Ireland
Duration: Sep 9 2012Sep 13 2012

Other

Other6th ACM Conference on Recommender Systems, RecSys 2012
CountryIreland
CityDublin
Period9/9/129/13/12

Fingerprint

Mean square error
Factorization
Feedback

Keywords

  • Matrix factorization
  • Recommender system
  • Social network

ASJC Scopus subject areas

  • Software

Cite this

Yang, X., Steck, H., Guo, Y., & Liu, Y. (2012). On top-k recommendation using social networks. In RecSys'12 - Proceedings of the 6th ACM Conference on Recommender Systems (pp. 67-74) https://doi.org/10.1145/2365952.2365969

On top-k recommendation using social networks. / Yang, Xiwang; Steck, Harald; Guo, Yang; Liu, Yong.

RecSys'12 - Proceedings of the 6th ACM Conference on Recommender Systems. 2012. p. 67-74.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Yang, X, Steck, H, Guo, Y & Liu, Y 2012, On top-k recommendation using social networks. in RecSys'12 - Proceedings of the 6th ACM Conference on Recommender Systems. pp. 67-74, 6th ACM Conference on Recommender Systems, RecSys 2012, Dublin, Ireland, 9/9/12. https://doi.org/10.1145/2365952.2365969
Yang X, Steck H, Guo Y, Liu Y. On top-k recommendation using social networks. In RecSys'12 - Proceedings of the 6th ACM Conference on Recommender Systems. 2012. p. 67-74 https://doi.org/10.1145/2365952.2365969
Yang, Xiwang ; Steck, Harald ; Guo, Yang ; Liu, Yong. / On top-k recommendation using social networks. RecSys'12 - Proceedings of the 6th ACM Conference on Recommender Systems. 2012. pp. 67-74
@inproceedings{79f8048c087b4ec09c163db358e5e95d,
title = "On top-k recommendation using social networks",
abstract = "Recommendation accuracy can be improved by incorporating trust relationships derived from social networks. Most recent work on social network based recommendation is focused on minimizing the root mean square error (RMSE). Social network based top-k recommendation, which recommends to a user a small number of items at a time, is not well studied. In this paper, we conduct a comprehensive study on improving the accuracy of top-k recommendation using social networks. We first show that the existing social-trust enhanced Matrix Factorization (MF) models can be tailored for top-k recommendation by including observed and missing ratings in their training objective functions. We also propose a Nearest Neighbor (NN) based top-k recommendation method that combines users' neighborhoods in the trust network with their neighborhoods in the latent feature space. Experimental results on two publicly available datasets show that social networks can significantly improve the top-k hit ratio, especially for cold start users. Surprisingly, we also found that the technical approach for combining feedback data (e.g. ratings) with social network information that works best for minimizing RMSE works poorly for maximizing the hit ratio, and vice versa.",
keywords = "Matrix factorization, Recommender system, Social network",
author = "Xiwang Yang and Harald Steck and Yang Guo and Yong Liu",
year = "2012",
doi = "10.1145/2365952.2365969",
language = "English (US)",
isbn = "9781450312707",
pages = "67--74",
booktitle = "RecSys'12 - Proceedings of the 6th ACM Conference on Recommender Systems",

}

TY - GEN

T1 - On top-k recommendation using social networks

AU - Yang, Xiwang

AU - Steck, Harald

AU - Guo, Yang

AU - Liu, Yong

PY - 2012

Y1 - 2012

N2 - Recommendation accuracy can be improved by incorporating trust relationships derived from social networks. Most recent work on social network based recommendation is focused on minimizing the root mean square error (RMSE). Social network based top-k recommendation, which recommends to a user a small number of items at a time, is not well studied. In this paper, we conduct a comprehensive study on improving the accuracy of top-k recommendation using social networks. We first show that the existing social-trust enhanced Matrix Factorization (MF) models can be tailored for top-k recommendation by including observed and missing ratings in their training objective functions. We also propose a Nearest Neighbor (NN) based top-k recommendation method that combines users' neighborhoods in the trust network with their neighborhoods in the latent feature space. Experimental results on two publicly available datasets show that social networks can significantly improve the top-k hit ratio, especially for cold start users. Surprisingly, we also found that the technical approach for combining feedback data (e.g. ratings) with social network information that works best for minimizing RMSE works poorly for maximizing the hit ratio, and vice versa.

AB - Recommendation accuracy can be improved by incorporating trust relationships derived from social networks. Most recent work on social network based recommendation is focused on minimizing the root mean square error (RMSE). Social network based top-k recommendation, which recommends to a user a small number of items at a time, is not well studied. In this paper, we conduct a comprehensive study on improving the accuracy of top-k recommendation using social networks. We first show that the existing social-trust enhanced Matrix Factorization (MF) models can be tailored for top-k recommendation by including observed and missing ratings in their training objective functions. We also propose a Nearest Neighbor (NN) based top-k recommendation method that combines users' neighborhoods in the trust network with their neighborhoods in the latent feature space. Experimental results on two publicly available datasets show that social networks can significantly improve the top-k hit ratio, especially for cold start users. Surprisingly, we also found that the technical approach for combining feedback data (e.g. ratings) with social network information that works best for minimizing RMSE works poorly for maximizing the hit ratio, and vice versa.

KW - Matrix factorization

KW - Recommender system

KW - Social network

UR - http://www.scopus.com/inward/record.url?scp=84867381268&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84867381268&partnerID=8YFLogxK

U2 - 10.1145/2365952.2365969

DO - 10.1145/2365952.2365969

M3 - Conference contribution

SN - 9781450312707

SP - 67

EP - 74

BT - RecSys'12 - Proceedings of the 6th ACM Conference on Recommender Systems

ER -